COFFEE CONSTITUENTS AND MODULATION OF

ANTIOXIDANT STATUS IN CACO2 CELLS

by

YAZHENG LIU

B. Sc., Shandong University, 2006

A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE

in

THE FACULTY OF GRADUATE STUDIES (Food Science)

THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)

March 2010

© Yazheng Liu 2010 ABSTRACT

Coffee contains biologically active components which may affect chronic disease risk.

These biologically active components include , cafestol and kahweol, and antioxidants such as chlorogenic acids and Maillard reaction products (MRPs) that are generated during roasting. Although MRPs are regarded as being the most abundant group of antioxidants present in coffee, the mechanism underlying the antioxidant effects of coffee MRPs in both in vitro and in biological systems has yet to be elucidated.

In this study, the in vitro antioxidant properties of roasted and nonroasted coffee extracts

(Coffea arabica L.) were tested using oxygen radical absorbance capacity (ORAC),

Trolox equivalent antioxidant capacity (TEAC) and reducing power assays. MRPs were shown to be the prevailing antioxidants in roasted coffee extracts. The mechanisms of the antioxidant action associated with coffee MRPs involve the hydrogen atom transfer (HAT) mechanism and the single electron transfer (SET) mechanism.

The biological effects of MRPs derived from coffee extracts on the enzymatic antioxidant defense in human colon adenocarcinoma Caco2 cells were also investigated. No induction of antioxidant activities of catalase, glutathione peroxidase, glutathione reductase and were observed in Caco2 cells after exposure to coffee MRPs, except for an increased glutathione peroxidase activity after 24 h exposure.

In contrast, significantly decreased activities of catalase and glutathione peroxidase, and a reduced glutathione content were observed in Caco2 cells after treatment with coffee

MRPs (p<0.05).

ii The antioxidant gene expression profile in Caco2 cells after coffee treatment was further investigated using a RealTime Polymerase Chain Reaction (PCR) array technology.

Results demonstrated that roasted coffee extracts induced the expression of specific antioxidant response element (ARE)driven genes in Caco2 cells, thus enhancing cellular endogenous defense systems. This is the first report of the molecular mechanism underlying the antioxidant effect of coffee in Caco2 cells. Hydrogen peroxide generated in the cell culture system as a consequence of coffee exposure, may serve as a signaling molecule that is involved in the gene regulatory effect associated with coffee extracts.

iii TABLE OF CONTENTS

ABSTRACT...... ii

TABLE OF CONTENTS...... iv

LIST OF TABLES...... viii

LIST OF FIGURES ...... x

LIST OF ABBREVIATIONS...... xiii

ACKNOWLEDGEMENTS...... xvi

CHAPTER I

OVERVIEW: GENERAL INTRODUCTION, LITERATURE REVIEW, AND

RESEARCH HYPOTHESES AND OBJECTIVES...... 1

1.1 GENERAL INTRODUCTION...... 2

1.2 LITERATURE REVIEW ...... 6

1.2.1 Oxidative stress and antioxidants...... 6

1.2.1.1 Reactive oxygen species and oxidative stress...... 6

1.2.1.2 Antioxidant mechanisms...... 7

1.2.1.3 Gene regulations by oxidative stress...... 7

1.2.2 Maillard Reaction (MR)...... 10

1.2.2.1 Chemistry of Maillard Reaction (MR)...... 10

1.2.2.2 Chemistry of Maillard reaction products (MRPs) ...... 11

1.2.2.3 Antioxidant properties of MRPs ...... 14

1.2.2.4 MRPs and chemoprotective ...... 17

1.2.3 Coffee – a source of MRPs ...... 18

1.2.3.1 Composition of coffee bioactive components...... 19

iv 1.2.3.2 Coffee as a source of dietary antioxidant...... 25

1.2.3.3 Coffee consumption and health ...... 29

1.3 RESEARCH HYPOTHESES AND OBJECTIVES...... 32

CHAPTER II

CHEMICAL CHARACTERISTERCS AND ANTIOXIDANT PROPERTIES OF

COFFEE...... 35

2.1 Introduction...... 36

2.2 Materials and methods ...... 38

2.2.1 Coffee preparation ...... 38

2.2.2 Ultrafiltration ...... 38

2.2.3 Physical chemical analyses ...... 39

2.2.4 Chemical based antioxidant assays...... 40

2.2.5 Cell based assays...... 42

2.2.6 Statistical analysis...... 47

2. 3 Results...... 48

2. 3.1 Yields and recovery of coffee brews and ultrafiltration fractions ...... 48

2.3.2 Chemical characteristics of coffee...... 50

2.3.3 Antioxidant activity of coffee extracts in chemical systems ...... 55

2.3.4 Biological effects of coffee extracts ...... 60

2.4 Discussion...... 66

2.4.1 Chemical characteristics of coffee...... 66

2.4.2 Antioxidant activity of coffee extracts...... 68

2.4.3 Biological effects of coffee extracts ...... 70

v CHAPTER III

COFFEE CONSTITUENTS AND MODULATION OF OXIDATIVE STATUS IN

CACO2 CELLS...... 72

3.1 Introduction...... 73

3.2 Materials and method...... 75

3.2.1 Preparation of coffee and Maillard reaction products (MRPs)...... 75

3.2.2 Chemical analyses and Antioxidant assays...... 76

3.2.3 Cellular in vitro Assay ...... 77

3.2.4 RealTime Quantitative Reverse Transcription PCR (RQ RTPCR) Array .... 78

3.2.5 Statistical analysis...... 81

3.3 Results...... 82

3.3.1 Recovery of coffee brews and fractions...... 82

3.3.2 Chemical characteristics of coffee...... 83

3.3.3 Antioxidant activity of roasted and green coffee...... 88

3.3.4 Biological effects of coffee bean extracts...... 91

3.3.5 Biological effects of coffee fractions on Caco2 cells ...... 99

3.3.6 The regulatory effects of coffee on the expression of the genes involved in the

oxidative stress and antioxidant defense system in Caco2 cells...... 105

3.3.7 Chemical characteristics and antioxidant activity of SucSer and AraSer

model MRPs...... 109

3.3.8 Chemical characteristics and antioxidant activity AraSer model MRPs

fractions...... 117

3.3.10 Cellbased bioactivity of AraSer MRPs ...... 121

vi 3.3.11 Biological effects of AraSer MRPs fractions on Caco2 cells ...... 128

3.3.12 Gene regulation of MRPs on the human oxidative stress and antioxidant

defense system (HOSAD) in Caco2 cells...... 132

3.4 Discussion...... 135

3.4.1 Chemical characteristics of coffee extracts and model MRPs...... 135

3.4.2 Antioxidant activity and reducing power of coffee constituents ...... 138

3.4.2 Biological effects of green coffee, roasted coffee and model MRPs on the

antioxidant defense system in Caco2 cells ...... 143

3.4.3 Coffee and the expression of Redoxsensitive genes in Caco2 cells...... 147

CHAPTER IV

GENERAL DISCUSSION AND CONCLUSIONS...... 155

4.1 GENERAL DISCUSSION ...... 156

4.1.1 Chemical characteristics and antioxidant properties of coffee ...... 156

4.1.2 Coffee, antioxidant enzymes and antioxidant genes...... 157

4.2 CONCLUSION...... 159

4.3 SUGGESTIONS FOR FUTURE RESEARCH...... 161

REFERENCES ...... 162

APPENDIX...... 185

vii LIST OF TABLES

Table 1.1 Example of antioxidant defense systems……………………………………….9

Table 1.2 Composition of green and roasted coffee……………………………………..19

Table 1.3 Caffeine content of different coffee beverages…………………….…………20

Table 1.5 Summary of potential health benefits of coffee consumption from

epidemiological studies…………………………………….…….…………...31

Table 2.1 Recovery yields of coffee extracts………………………………………….49

Table 2.2 Recovery of coffee fractions by water and salt ultrafiltrations………………..49

Table 2.3 Color parameters (L, E) and browning of fractionated coffee extracts…..51

Table 2.4 Antioxidant activities of coffee extracts and fractions……………………..…57

Table 2.5 Antioxidant activity of defatted nonfractionated coffee extracts and

recombined extracts…………………………………………………..….…..58

Table 2.6 Antioxidant activities of coffee fractions by water and salt ultrafiltration……59

Table 2.7 IC 50 of coffee extracts on Caco2 cells using MTT assay……...……………..62

Table 3.1 Human oxidative stress and antioxidant defense PCR array gene table ……...79

Table 3.2 Recovery of coffee fractions by ultrafiltration……………………..….………82

Table 3.3 Lightness (L) and browning of coffee extracts and untrafiltration fraction.…83

Table 3.4 Antioxidant activity of coffee extracts…………………..……………………89

Table 3.5 Antioxidant activity of coffee fractions determined by the ORAC method….90

viii Table 3.6 Antioxidant activity of coffee fractions determined by the TEAC method…..91

Table 3.7 Antioxidant activity of coffee fractions determined by the RP method……91

Table 3.9 IC 50 of coffee extracts on Caco2 cells using MTT assay…………………….92

Table 3.10 Caco2 MTT response to coffee with and without H 2O2 treatment………101

Table 3.11 Genes differently expressed in Caco2 cells after incubation with coffee

extracts and hydrogen peroxide...……………………………………….…107

Table 3.12 Lightness (L) and browning of model MRPs………………..…..…….111

Table 3.13 Antioxidant activity of SucSer and AraSer MRPs……………..……….116

Table 3.14 Recovery of AraSer MRPs ultrafiltration fractions………………………..117

Table 3.15 Lightness (L) and browning of fractions derived from AraSer model MR

system……………………………….…………………..……………....….118

Table 3.16 Antioxidant activity of fractionated AraSer MRPs…………….………….121

Table 3.17 IC 50 values of AraSer MRP extracts on Caco2 cells using MTT assay….124

Table 3.18 Caco2 MTT response to MRPs with (+) and without () H 2O2 treatment…128

Table 3.19 Genes differently expressed in Caco2 after incubation with AraSer MRP.133

ix LIST OF FIGURES

Figure 1.1 CGA degradation during roasting of coffee bean………………..……...……22

Figure 1.2 Contribution of coffee to the antioxidant intake in diet……………………....26

Figure 1.3 Mechanism of scavenging of free radicals by caffeine………………...…….29

Figure 2.1 Fluorescence emission spectra (350550 nm) of light roasted and dark roasted

coffee extracts and fractions.……………………………...……………..….53

Figure 2.2 Comparison of the UVvisible spectra of light roasted and dark roasted coffee

extracts and fractions…….…………………………………………….……54

Figure 2.3 Effects of light roasted and dark roasted coffee extracts on the tetrazolium

reduction rate in the MTT assay…………………….……………..………..61

Figure 2.4 Effect of coffee extracts on catalase (CAT) activity in Caco2 cells………...63

Figure 2.5 Effect of coffee extracts on glutathione (GSH) content in Caco2 cells……..65

Figure 3.1 Fluorescence emission spectra (350550 nm) of green bean, light roasted and

dark roasted coffee extracts and fractions.………………………..…….…..87

Figure 3.2 Comparison of the UVvisible spectra of green bean, light roasted and dark

roasted coffee extracts and fractions……………………………….….…....88

Figure 3.3 Alphadicarbonyl compounds in light roasted and dark roasted coffee

extracts………………………….………………………………….….…....89

Figure 3.4 Effects of green bean, dark roasted and light roasted coffee extracts on the

tetrazolium reduction rate in the MTT assay……………………...……..….94

Figure 3.5 Effect of coffee extracts on glutathione (GSH) content in Caco2 cells……..95

Figure 3.6 Effect of coffee extracts on glutathione peroxidase (GPX) activity in Caco2

cells……………………….…………………………………………….…...97

x Figure 3.7 Effect of coffee extracts on catalase (CAT) activity in Caco2 cells….…….98

Figure 3.8 Effect of light roasted and green bean coffee extracts on Caco2 cellular GSH

contents with and without H 2O2 treatment……………….....……………..101

Figure 3.9 Effect of light roasted and green bean coffee extracts on Caco2 cellular

antioxidant enzyme activities with and without H 2O2 treatment……….….104

Figure 3.10 Antioxidant genes expression in Caco2 cells treated with light roasted, dark

roasted coffee extracts and H 2O2 compared to those in control cells.…..…108

Figure 3.11 Fluorescence emission spectra (350550 nm) of light roasted and dark roasted

SugarSerine MRPs extracts………………...………………………..……111

Figure 3.12 UV spectra of light roasted and dark roasted SugarSerine MRPs extract...111

Figure 3.13 Alphadicarbonyl compounds in AraSer MRPs and SucSer MRPs crude

extracts…………………………………….……………………………….115

Figure 3.14 Fluorescence emission spectra (350550 nm) of light roasted and dark roasted

MRPs extracts and fractions……………………….………………………119

Figure 3.15 Comparison of the UV spectra of light roasted and dark roasted MRPs

extracts and fractions………………………………………………………120

Figure 3.16 Effects of light roasted and dark roasted) AraSer MRPs extracts on the

tetrazolium reduction rate in the MTT assay………………………………123

Figure 3.17 Effect of AraSer MRPs extracts on glutathione (GSH) content in Caco2

cells...... 125

Figure 3.18 Effect of AraSer MRPs extracts on glutathione peroxidase (GPX) activity in

Caco2 cells……………………………………………………………...…126

xi Figure 3.19 Effect of AraSer MRPs extracts on superoxide dismutase (SOD) activity in

Caco2 cells……………………………………………………….………..127

Figure 3.20 Effect of MRPs extracts and associated fractions derived from light roasted

AraSer MR system on Caco2 cellular glutathione (GSH) contents after 24 h

of treatment……………………………………………………….…..……129

Figure 3.21 Effect of light roasted (LR) AraSer MRP extracts and fractions on Caco2

cellular antioxidant enzyme activities after 24 h of treatment…...……..….130

Figure 3.22 Antioxidant genes expression in Caco2 cells treated with light roasted and

dark roasted AraSer MRPs compared to those in control cells……….…..134

Figure 4.1 UVvis spectrum of a typical melanoidin and of individual chromophoric sub

structures…………..……………………………………………..……...…137

Figure 4.2 Chlorogenic acids and related compounds according to chemical

characteristics…………………………………………………………..…..140

Figure 4.3 Proposed pathway of H 2O2 formation in coffee brews……………………...154

xii LIST OF ABBREVIATIONS

AAPH 2,2’azobis(2amidinopropane) dihydrochloride

ABTS 2,2’azinobis(3ethylbenzothiazoline 6sulfonate)

AP1 Activator protein1

Ara Arabinose

ARE Antioxidant response element

BHA Butylated hydroxyanisole

BHT Butylated hydroxytoluene

C+K Cafestol and kahweol

Caco2 Human intestinal adenocarcinoma cancer cell line

CAT Catalase

CCR NADPHcytochrome creductase

CFAQ Caffeoylferuloylquinic acids

CGA Chlorogenic acid

CQA Caffeoylquinic acid

DR Dark roasted

GB Green coffee beans

GPX Glutathione peroxidase

GR Glutathione reductase

GSH Glutathione

GST GlutathioneS h Hours

HAT Hydrogen atom transfer

xiii HepG2 Human liver carcinoma cell line

HMF Hydroxymethylfurfural

HPLC High performance liquid chromatography iNOS Inducible nitric oxide synthase

Int407 Human embryonic intestinal cell line

Keap1 Kelchlike ECHassociated protein 1

LR Light roasted

MEM Minimum essential medium

MR Maillard reaction

MRPs Maillard reaction products

MTT 3(4,5dimethylthiazol2yl) 2,5diphenyltetrazolium bromide

MW Molecular weight

NFκB Nuclear factor kappa B

Nrf2 Nuclear factorerythroid 2 p45 subunitrelated factor 2

• O Superoxide radicals 2

ONOO Peroxynitrite radical

ORAC Oxygen radical absorbance capacity

PBS Phosphate buffer saline pCoA pcoumaric acid

PCR Polymerase chain reaction

PR Reducing power

• RO Peroxyl radicals 2

xiv ROS Reactive oxygen species

SD Standard deviation

Ser Serine

SET Single electron transfer

SOD Superoxide dismutase

Suc Sucrose tBOOH tert butylhydroperoxide

TE Trolox equivalents

TEAC Trolox equivalent antioxidant capacity

xv ACKNOWLEDGEMENTS

My graduate study was a journey with my supervisor and committee, my parents, my grandparents, lab colleagues, graduate secretaries and many friends from UBC Bible study group. They are the people who make this journey colorful, enjoyable and unforgettable, and finally to fruition.

My supervisor, Dr. David Kitts, has a special way to keep me going. He told jokes to encourage me when I fell down, and treated me for coffee when I accomplished even little things. He was always willing to help me, and he never pushed me too much. He said: when you are happy, I am happy. I would like to thank Dr. Kitts for his care, guidance, patience and trust. Especially, I am very thankful for the countless hours he has spend helping me with the scientific writing. Dr. Scaman always has a smile on her face when she saw me. I can feel the encouragement and care through the smile. I appreciate her time to answer my questions regarding to the statistical analysis, and especially thank her for the concern and understanding of the progress of my graduate study. I want to give special thanks to Dr. Adams for allowing me to use the equipment in his lab. I also want to thank him for making the time to my meetings and replying my emails very quick.

There were sunny days and rainy days during the journey. However, no matter what kind of day, my grandma and my parents were with me and supported me with all they could.

Because of them, I was able to run through the whole journey. Thank you very much for your prayers! Grandma! Thank you very much! Dad! Mom! I love you! I am heartily thankful to my grandpa, Don Smith, and grandma, Elaine Smith, whose encouragement, love and support from the beginning to the last, enabled me to stay in this foreign country with joy and happiness. My grandma is a good listener and advisor. I shared many things

xvi with her, things that I was happy and unhappy with. I would like to thank grandma for listening to my problems and always staying on my side. I would like to extend great thanks to Shaowei Dong for his constant love and support, and those delicious dishes that he made for me. I would also like to give special thanks to Andrea Goldson, who helped me with research problems, revise my writings, and took me to wonderful events and activities, which are all good memories of my graduate study. She did so much for me. I would like to give my deepest thanks to Andrea. I am also appreciative of Fanchui Gang, who taught me how to use Endnote and shared his knowledge of chemistry. I am grateful for working in Dr. Kitts’ lab with an excellent team of graduate students and lab technicians. I would like to thank Ingrid Elisia, Xiumin Chen, Alexandra Tijerina Saenz,

Monica Purnama, Steve Tomiuk, Minh Huynh and Katie Hu for teaching me techniques, helping me with research problems and giving me guidance for graduation. Also, I would like to thank Kirsten Cameron, Lia Maria Dragan, Allison Barnes, Val Skura and Pedro

Aloise for their help, support and hard work. This thesis would not have been possible without the prayers from UBC Bible study group. I would like to give sincere thanks for their prayers. Lastly, I would like to give my regards and blessings to all of those who supported me in any respect during the completion of my project.

xvii

CHAPTER I

OVERVIEW: GENERAL INTRODUCTION, LITERATURE REVIEW, AND

RESEARCH HYPOTHESES AND OBJECTIVES

1 1.1 GENERAL INTRODUCTION

The role of food and its components on human health continues to be a keen topic of

interest to both the lay public and academics. The potential of forming beneficial or

harmful products in food as a result of food processing is a major area of investigation by

food toxicologists. One particular reaction that has received continuing interest for more

than 100 years is the Maillard, or browning reaction. This reaction takes place during the

thermal processing of food and involves the condensation of amino groups from amines,

amino acids, peptides, or proteins with carbonyl groups from sugars or fatty acids

(Maillard, 1912; Hodge, 1953). The reaction develops into a complex network of

chemical intermediate products, including fluorescent, color and flavor compounds, and polymerized brown endproducts, all herein referred to as Maillard reaction products

(MRPs) (Hodge, 1953). MRPs can be divided into two classes; namely the low molecular

weight colored compounds that consist of four linked ring structures, with molecular

weights below 1 KDa, and the high molecular weight MRPs, which are colored polyphenolic structured compounds (Arnoldi et al. , 1997; Hofmann, 1997, 1998a; Ames et al. , 1999a). The high molecular weight MRPs, termed melanoidins, have a molecular weight up to 300 KDa (Ibarz et al. , 2009).

Besides the sensory properties attributed to MRPs, some negative aspects of this reaction

also exist, including the destruction of essential amino acids, a decrease in digestibility of proteins, and the potential production of products with carcinogenic, mutagenic and toxic potential; all of which can have an impact on both the quality and safety of heated foods

(OBrien and Morrissey, 1989). On the other hand, however, some MRPs have been

2 demonstrated to have healthpromoting properties. Melanoidins, for example, recently

have been shown to possess antioxidant activity (Wijewickreme and Kitts, 1997, 1998a;

DelgadoAndrade et al. , 2005), antimicrobial activity (RufianHenares and Morales,

2007a), antihypertensive activity (RufianHenares and Morales, 2007c), chemopreventive and antimutagenic properties (Powrie et al. , 1986; Faist et al. , 2001; Somoza et al. , 2003),

as well as prebiotic effects (Ames et al. , 1999b).

MRPs are an abundant group of compounds that also exist in coffee brews and therefore

represent a significant part of the diet for those that consume coffee beverages. High

molecular weight melanoidins (MW>10KDa) derived from coffee have antioxidant properties that involve metal chelating of prooxidants (Wijewickreme and Kitts, 1998b;

Morales et al. , 2005), free radical scavenging of peroxyl radicals and hydroxyl radicals

(Morales and JimenezPerez, 2004; Morales, 2005), and other free radicals (Borrelli et al. ,

2002). In line with the chemical antioxidant activity, coffee melanoidins have also been

shown to protect human hepatoma HepG2 cells against oxidative damage induced by

tert butylhydroperoxide (Goya et al. , 2007). However, a dose dependent reduction of

glutathione (GSH) content in HepG2 cells was also observed after 24 h coffee

melanoidins treatment (Goya et al. , 2007). The mechanism underlying this observation is

unknown. The question remains as to whether MRPs that show antioxidant activity in

chemical assays, will also function as antioxidants in biological systems.

Relatively little research has been dedicated to study the antioxidant properties of low

molecular weight MRPs (MW<1KDa), derived from coffee brews. One study found that

low molecular weight coffee components (MW<1KDa) possess higher antioxidant

activity compared to the high molecular weight components (Somoza et al. , 2003), which

3 indicates that the low molecular weight MRPs derived from coffee may have higher antioxidant activity compared to the high molecular weight melanoidins. It is still unclear

if there is a relationship between the antioxidant activity of MRPs and the related

molecular weights.

Caco2 cells derived from human colon adenocarcinoma cells are widely used to model

the small intestine for investigating the effects of food components on cell function and

metabolism. These cells are often used to examine the biological response to an oxidative

stress and mitigation by nutritional antioxidants (Baker and Baker, 1993; Cepinskas et al. ,

1994; Manna et al. , 1997; Wijeratne et al. , 2005). The activity and genetic expression of

antioxidant enzymes in Caco2 cells have been reported to change with time after

confluence (Baker and Baker, 1992). Therefore, both homogeneously undifferentiated (at

subconfluence) and 100% confluence Caco2 cells were used in this thesis to study the

oxidative stress responses induced by both coffee and model MRPs.

This research thesis is composed of two experiments. Experiment I attempted to

characterize and compare the antioxidant properties of light roasted and dark roasted

coffee, and related fractions that have different molecular weights. The biological effects

of coffee extracts on the cellular antioxidant status in Caco2 cells were also investigated.

The objectives for Experiment II were to evaluate the antioxidant activity of MRPs

derived from coffee brews and to further explore the underlying antioxidant mechanisms

associated with coffee MRPs. The biological effects of MRPs derived from coffee on the

cellular antioxidant defense in Caco2 cells were also investigated. Finally, human

oxidative stress and antioxidant defense PCR array was used to gain an understanding of

changes in the antioxidant gene expression in Caco2 cells after coffee extract treatment.

4 This experiment provides an insight into the mechanisms of antioxidant status modulatory effects associated with coffee constituents at the molecular level.

5 1.2 LITERATURE REVIEW

1.2.1 Oxidative stress and antioxidants

1.2.1.1 Reactive oxygen species and oxidative stress

Oxygen is an essential requirement for normal growth and metabolism of the body.

However, related products of cellular respiration, often referred to as reactive oxygen species (ROS), possess a tremendous potential for toxicity which is manifested by

oxidation of important cellular constituents that can eventually result in cell death. ROS

are generated endogenously by autoxidation and through cellular metabolic reactions that involve both mitochondria and peroxisomes, not to mention, many cytosolic enzymes and

nonenzymatic systems (Sies, 1993). In addition, exposure to exogenous factors, such as

ultraviolet light, radiation, chemotherapeutics, smoking and some specific environmental toxins will also trigger the production of ROS (Leanderson and Tagesson, 1990; Sies,

1993). ROS attack lipids, exciting chain reactions that can cause cumulative oxidative damage. Since many studies indicate ROS to be an underlying cause for ageing, chronic disease and death (Raha and Robinson, 2000), there is a continuous requirement for the inactivation of ROS within the body. High doses and/or inadequate removal of ROS will result in oxidative stress (Sies and Cadenas, 1985), a condition which has been proposed to initiate mutagenesis, carcinogenesis and cardiovascular disease (Waris and Ahsan,

2006; Valko et al. , 2007).

6 1.2.1.2 Antioxidant mechanisms

The exposure of cells and organ systems to a high partial pressureoxygen environment

will result in oxidative stress. Survival from a hyperbaric state is possible through the

action of strategically located enzymatic and nonenzymatic antioxidants, and by the continued replacement and repair of oxidative damaged tissue macromolecules. The term

“antioxidant” can therefore be used to describe any substance that delays or inhibits oxidative reactions, albeit the ultimate effectiveness on removing oxidative stress will differ (Kitts, 1997).

1.2.1.2.1 Nonenzymatic and enzymatic antioxidants

A cellular antioxidant defense system consists of a collective function of nonenzymatic and enzymatic antioxidants that work in concert to reduce the potential toxicity of ROS.

Table 1.1 provides an overview of some antioxidants that have received interest from scientists and characterizes specific antioxidant defense mechanisms (Sies, 1993; Yuan and Kitts, 1997). These include: (1) the scavenging of free radicals and singlet oxygen

(e.g. vitamin E and superoxide dismutase), (2) the reduction of hydroperoxides (e.g. glutathione peroxidase and catalase), (3) the removal of metal catalysts (prooxidants) from the site of action (e.g. proteins and chelating agents).

1.2.1.3 Gene regulations by oxidative stress

The ability of cells to cope with, or prevent, the damage caused by oxidative stress is an

important component of cellular, and by extension whole body homeostasis. In human

cells, gene expression is modulated by ROS (Valko et al. , 2007). This modulation has been observed in response to both direct and indirect oxidative challenge and involves

7 changes at many levels that include transcription, mRNA stability, and signal

transduction (Crawford et al. , 1988; Devary et al. , 1991; Wang et al. , 1996; Hartsfield et al. , 1997). Numerous specific genes related to the enzymatic antioxidant defense system have been identified. These include genes that encode glutathione peroxidases (GPX), peroxiredoxins (PRDX), superoxide dismutases (SOD), and oxidative stress responsive genes, not to mention other genes that are involved in ROS metabolism. Modest inductions by oxidative stress of enzymatic antioxidant mechanisms, such as SOD, GPX, or catalase (CAT), have been observed in mouse muscle cells (Franco et al. , 1999).

Studies have found that several protooncogenes, for example, cfos , cmyc, and cjun , were induced by ROS, such as hydrogen peroxide (Crawford et al. , 1988; Devary et al. ,

1991; Nose et al. , 1991). These genes are critical to cellular growth and differentiation and an onset of an aberrant expression may potentially result in cancer (Zheng and

Hendry, 1997). Oxidant/antioxidant responsive elements (ARE) have been identified in the promoter region of several genes, which include glutathione S transferase (GST) Ya subunit, cfos , cjun , heme oxygenase and NAD(P)H:quinine (Rushmore and Pickett, 1993; Venugopal and Jaiswal, 1998). These genes are stimulated at the transcriptional level by hydrogen peroxide. Reports have identified an element in the

GST Ya subunit that is bound and activated by a series of antioxidant compounds

(Nguyen et al. , 1994). Interestingly, the affinity to stimulate transcription by antioxidants is based on how well they produce critical levels of ROS (Crawford, 2002). Several important transcriptional factors have been identified that are mediators of oxidative stress (Hancock et al. , 2001). These factors are induced, or activated, by ROS and then bind and activate those genes that are involved in the overall cellular antioxidant defense

8 systems. Most notably, these include nuclear factorkappa B (NFκB) and activator protein1 (AP1). These genes are also excellent biomarkers for assessing cellular antioxidant defense systems, and represent valuable potential targets in the treatment of oxidative stress related diseases.

Table 1.1 Example of antioxidant defense systems (Sies, 1993; Yuan and Kitts, 1997) System Functionality Nonenzymatic Albumin binds Fe, Cu ions Ascorbate (vitamin C) electron donor , singlet oxygen quencher, regenerate α tocopherol radical Bilirubin plasma antioxidant Flavonoids plant antioxidants Glutathione (GSH) thiol group maintain redox potential Lycopene electron donor , singlet oxygen quencher Ubiquinol10 radical scavenger Urate radical scavenger Uric acid free iron binding αtocopherol (vitamin E) radical chainbreaking: electron donor, hydrogen donor, free radical scavenger, singlet oxygen quencher βcarotene electron donor , singlet oxygen quencher Enzymatic (direct) Catalase (CAT) mainly located in cellular peroxisomes and to some in the cytosol; catalyzes the reduction of hydrogen peroxide. GSH peroxidases (GPX) plasma, intracellular. Reduce hydrogen peroxide and lipid peroxides to water and lipid alcohols. Superoxide dismutases (SOD) plasma, milk, cytosol, mitochondria. Contain redox metals in the catalytic center and convert dismutase superoxide radicals to hydrogen peroxide and oxygen. Enzymatic (ancillary enzymes) Conjugation enzymes glutathionestransferase, UDPglucuronosyl : conjugates xenobiotics and alkylating agents with GSH GSSG reductase maintain GSH levels NADPH supply NADPH for GSSG reductase NADPHquinoe oxidoreductase twoelectron reduction Repair systems DNA repair systems oxidized protein turnover oxidized phospholipid turnover Transport systems GSSG export Thioether (sconjugate) export

9 1.2.2 Maillard Reaction (MR)

The Maillard reaction (MR) occurs quickly during heating and was first reported by L.C.

Maillard in 1912. It is a complex series of nonenzymatic reactions that involves free

amino groups reacting with carbonyl groups that result in a browning reaction. MR is one

example of a nonenzymatic browning reaction that is very important in many food

systems and produces desirable attributes such as flavour, texture and color of food.

Many intermediate products with bioactive properties are also generated during the

Maillard reaction. These may include potential carcinogens, mutagens, antimutagens,

antioxidants, allergens and antiallergens (Friedman, 2005), albeit considerable

inconsistency in the scientific literature exists.

1.2.2.1 Chemistry of Maillard Reaction (MR)

The complete chemical description of the MR is yet to be fully defined. The earliest

systematic review of the reaction scheme was put forward by Hodge in 1953, and further

modified by Reynolds in 1969 and Mauron in 1981. Basically, the reaction was divided

into three stages which are:

(1) The initial stage, which consists of sugaramine condensation and Amadori or Heyns

rearrangement, forming a Schiff base.

(2) The intermediate stage, which consists of sugar dehydration and fragmentation, as

well as amino acid degradation. Many low molecular weight intermediate compounds are produced during this stage. For example, highly reactive αamino carbonyl compounds

are formed by Strecker degradation, the third reaction in the formation of MRPs. In this

reaction, condensation of these intermediate compounds produces heterocyclic

10 compounds, which contribute to many flavours in heated food, such as coffee. Some

fluorescent compounds and brown pigments also occur, but at very low concentrations at

these particular stage.

(3) The final stage of the MR is a polymerization reaction that produces high molecular

weight, colored endproducts. These are referred to specifically, as melanoidins in food systems and advanced glycation end products (AGEs) in body tissues.

1.2.2.2 Chemistry of Maillard reaction products (MRPs)

Maillard reaction products (MRPs) consist of a vast number of reaction products. In general, low molecular weight MRPs are very important in flavor and offflavor production, while high molecular weight MRPs/melanoidins are the ultimate end products of the reaction. Attempts to summarize the proposed structure of melanoidins

was made by Goya et al. (Goya et al. , 2007) and included different endproducts, such as:

(i) Low molecular weight colored substances that crosslink with free amino groups of

lysine or arginine in proteins; (ii) units of furan and/or pyrroles that react through polycondesation reactions to form melanoidin repeating units; (iii) the melanoidin

chemical skeleton, which is mainly built up from sugar degradation products formed in

the early stages of the reaction, and often polymerized and crosslinked by amino

compounds. Since the composition of MRPs is greatly influenced by several factors, such

as the ratio and type of amino compounds, together with the presence of reducing sugars, pH, temperature, time, and water activity (Wijewickreme and Kitts, 1997), it is expected

that the final composition of MRPs will reflect the complex reactions that are involved in

the multiple chemical schemes describing a variety of potential products.

11 Spectrophometric measurements are commonly used to characterize and quantitate the

generation of complex MRPs. Broad spectral peaks occurring between 250 nm to 350 nm

are often associated with low molecular weight MRPs (Jing and Kitts, 2003). Some

compounds derived from early stage MRPs generation, including pyrazines and

hydroxymethylfurfural (HMF) compounds, are detected near the maximum absorbance

at 280 nm (Lerici et al. , 1990; Ames et al. , 1999a). The alkaline degradation of reducing

sugars leads to the generation of chromophores at both 210 nm and 265 nm. Maximum

absorption at 265 nm has been attributed to the presence of αdicarbonyl intermediates

(Jing and Kitts, 2003). Polymerization of the intermediates occurs at the late stages of the

MR with the formation of melanoidins that are detectable in the visible region. A single

wavelength measurement of brown pigments at 420 nm is frequently used to measure the

rate and extent of the final stage of the Maillard reaction (Morales and JimenezPerez,

2004).

The color parameters provided by tristimulus colorimetry are also commonly applied to

indicate the visual color attributed to nonenzymatic browning (Morales and van Boekel,

1998; Morales and JimenezPerez, 2001). Chemical analyses of the brown pigments (or buff orange) has confirmed the presence of furans, pyrroles, and pyridines (Rizzi, 1997).

Pyrazine formation was related to color formation, and as pyrazine products increased the

color of the MRPs changed from colorless to yellow, then to brown and finally to darker brown pigments (Wong and Shibamoto, 1996). Morales and JimenezPerez (Morales and

JimenezPerez, 2001), examined the heating of model MRPs at 100 ºC for 24 h that

included glucosealanine (GA), glucoseglycine (GG), glucoselysine (GL), lactose

alanine (LA), glucoseglycine (LG), and lactoselysine (LL) and recorded changes in the

12 L, a, b tristimulus coordinates. The lightness indicator L decreased significantly during

heating in these model systems, which indicates increased darkness at the final stage of

the Maillard reaction. A net increase in a yellowbrown color was observed during the first hours of heating before reaching a maximum. This was followed by a color change to orangebrown, even purplishred for GL and LL MRP samples at higher heating times.

GL and LL systems produced the darkest color. High concentrations of sugars in the open chain form have been found to brown faster and more intensely (Boekel, 2001). Color formation is therefore due to both the presence of low molecular weight and the high molecular weight MRPs (Ames, 1992). Other workers suggested that the redness parameter, a, may be a reliable indicator of acrylamide levels, which are generated under very specific MR conditions and represent harmful intermediates in certain thermally processed foods, such as fried potatoes (Gokmen and Senyuva, 2006).

While the development of color is an important feature of the MR, some studies have placed emphasis on the generation of fluorescent MRPs. It has been suggested that

different MRPs with fluorescent properties are related to increased heating conditions,

such as at prolonged heating (Morales and JimenezPerez, 2001). Some fluorescent

MRPs are involved in the formation of colored MRPs and may be the precursors of brown pigments (Leclere and BirlouezAragon, 2001). The fluorescent compounds

mentioned above did not follow the same time dependent trend as colored compounds in

defining the MR model systems (Morales and JimenezPerez, 2001). Fluorescent

molecules are stable with prolonged heating, whereas complexes that are brown will

change almost linearly with the duration of heating time (Morales and JimenezPerez,

2001).

13 1.2.2.3 Antioxidant properties of MRPs

MPRs, especially melanoidins present in processed food and generated in model systems

have been intensively studied in recent years. Evidence for an antioxidant role for MRPs is supported by many in vitro and some in vivo studies (Mastrocola and Munari, 2000;

Faist and Erbersdobler, 2002; DelgadoAndrade et al. , 2005; Kitts and Hu, 2005;

Michalska et al. , 2008). Heating glucose with amino acids has resulted in a remarkable

scavenging activity towards hydroxyl radical (Kawane et al. , 1999). Several heterocyclic

MRPs, which are major flavour compounds, show antioxidant activity by inhibiting

hexanal oxidation and lipid peroxidation, and scavenging thyrosyl radicals (Macku and

Shibamoto, 1991). Melanoidins from the glucoseglycine model system exhibited

antioxidant properties by quenching ROS (Wagner et al. , 2002). It has also been reported that melanoidins prepared from a xyloseglycine model system have antioxidant activity comparable to BHA and BHT (Hayase et al. , 1999). Furthermore, MRPs from a glucose

tryptophan model system can exhibit a synergistic effect with tocopherol in inhibiting

lipid autoxidation (Chiu et al. , 1991). Studies indicate that the antioxidant activity of

MRPs is depended on the type of sugar (Wijewickreme and Kitts, 1997; Sun et al. , 2006;

Chen and Kitts, 2008b). It has been shown that the configuration of OH group on carbon

moieties C3 and C4 are important for the formation of MRPs and related antioxidant

activities (Sun et al. , 2006). MRPs can also exert prooxidant activities in some cases. For

example, glucoselysine model MRPs generate free radicals in the presence of trace

amounts of iron, which in turn causes the degradation of hyaluronan (Deguine et al. ,

1998). Other studies have shown that the fructoselysine model MRPs exhibit more

14 prooxidant and genotoxic activities compare to glucoselysine model MRPs in the presence of copper ions (Wijewickreme and Kitts, 1997).

The radical scavenging activity of MRPs can also progressively increase with the

intensity of heat treatment and the development of browning (Murakami et al. , 2002).

The antioxidant property of MRPs therefore occurs to some extent in the later stage of the

MR, or from the generation of melanoidins. However, other studies have reported that browning cannot be directly related to the free radical scavenging properties of MRPs

formed over prolonged heating conditions (Morales and JimenezPerez, 2001; Jing and

Kitts, 2002). For example, in some cases, fluorescence measurement of heated MR

system is correlated better with free radical scavenging activities (Morales and Jimenez

Perez, 2001). In MR model systems, where the development of melanoidins was the final

outcome, antioxidant activity of MRPs derived from the same sugar or same amino acid

model systems, was inversely related to the fluorescent intensity (Chen and Kitts, 2008b).

Some studies have also found that the high molecular weight MRPs, which contribute to

the color pigments, also show antioxidant activities (Monti et al. , 1999), while others

found the antioxidant activity occurred mainly within the intermediate and low molecular

weight MRPs (Nienaber and Eichner, 1995). It can be deduced that color and fluorescent properties of MRPs are useful indicators of the different stages of the Maillard reaction

and can be used as indicators of high molecular weight and low molecular weight MRPs

formation, but may not absolutely reflect antioxidant capacity potential.

The chemical antioxidant property of MRPs is well accepted and has important

applications to the food industry. However, the effects of MRPs on the antioxidant

enzyme activity have not always been found to be desirable. Former studies have

15 reported that glutathione reductase (GR) and catalase (CAT) activities, and glutathione

(GSH) content in human lymphocytes were decreased when exposed to MRPs derived

from different sugarlysine model systems (Yen et al. , 2002). The activities of GR, CAT

and ascorbate peroxidase (APX) decreased in mung bean seeds in proportion to the

increase of MRPs during storage (Murthy et al. , 2002). Sugarlysine MRPs inhibited the

antioxidant enzyme activity of superoxide dismutase (SOD), CAT, and glutathione peroxidase (GPX), and the total GSH content in human intestinal epithelial Caco2 cells, while sugarcasein model MRPs decreased SOD, GPX, GR activities in Int407 cells and had no effect on Caco2 cells (Jing and Kitts, 2004b). A study showed that the dicarbonyl compound, methylglyoxal, generated during the early stage of the Maillard reaction, can

inhibit GPX activity by binding to GSH binding sites (Park et al. , 2003). Feeding mice

methylglyoxal significantly decreased liver SOD, glutathioneStransferase (GST), CAT,

glyoxalase I and II antioxidant enzyme activities and was associated with a decrease in

GSH content, along with an increase in lipid peroxidation. It was suggested that

methylglyoxal generates free radicals, which in turn lowers the antioxidant status in

animals (Ueda et al. , 1998). Finally, one study found that a MRPrich diet, which had

antioxidant activity in vitro , had no effect on modifying the oxidative status in healthy

humans (Seiquer et al. , 2008). There is very little in vivo data to indicate an antioxidant

defense mechanism for MRPs in animals, and human intervention trials (Kitts et al. ,

1993). Thus the question remains whether dietary intake of these compounds can exert an

antioxidant effect in the human body beyond that of the gastrointestinal tract.

16 1.2.2.4 MRPs and chemoprotective enzymes

MRPs can also be recognized as xenobiotics; a term which classifies compounds that are

not formed endogenously and require detoxifying mechanisms to protect the organism

from harmful effects (Somoza, 2005). These detoxifying mechanisms collectively

contribute to a chemopreventive potential. Most chemopreventive, nonendogenously formed agents act by modulating the Phase I carcinogenactivating enzymes and Phase II detoxifying enzymes. Phase I metabolic transformations include reduction, oxidation, and hydrolytic reactions, while Phase II transformations act through conjugation reactions of the xenobiotics, or on Phase I metabolites. A decrease in Phase I enzyme activity and associated increase in Phase II enzyme activity are considered events that have the most effective chemopreventive potential (Somoza, 2005).

Phase I NADPHcytochrome creductase (CCR) and phase II GST in Caco2 cells have been shown to decrease after incubation with low molecular weight (<10 KDa) or high molecular weight (>10KDa) glucoseglysine melanoidins (Hofmann et al. , 2001). Similar

results were found for a glucosecasein model system, where GST activity in Caco2 cells

was decreased after exposure to nondialysed glucose/casein melanoidins (Faist, 2001).

These results also showed that the effects on CCR and GST were mediated by both high

molecular weight and low molecular weight MRPs. However, low molecular weight

compounds were more effective than high molecular weight compounds (Hofmann et al. ,

2001). Also, methyglyoxal, isolated from the low molecular weight MRPs fraction and

tested on Caco2 cells, was shown to increase the activity of CCR and decrease the

activity of GST (Hofmann et al. , 2001). In an animal study, mice fed a diet containing glucoselysine MRPs exhibited a significant decrease of Phase I hydrocarbon

17 hydroxylase (AHH) and Phase II UDP glucuronyltransferase (UDPGT) activities in the

small intestine mucosa (Kitts et al. , 1993). Generally, the chemopreventive effect of modeled MRPs in these studies was not overly promising. However, further studies were also carried out in order to see the effects of foodderived MRPs on Phase I and Phase II enzymes. The chemopreventive action of bread crust on cultured Caco2 cells resulted in an induction of GST and reduced CCR (Lindenmeier et al. , 2002). These results were

confirmed by animal feeding studies. When bread crust was fed to rats at a moderate,

human diet equivalent intake for 15 days, the activities of Phase II GST and UDPGT in

the liver increased, and the total antioxidant capacity in the plasma was also enhanced

(Somoza et al. , 2005). These studies also demonstrated that pronylated amino acids and proteins, as part of melanoidins can act as antioxidant and chemopreventive agents in

vitro and in vivo . Another chemopreventive compound formed during heattreatment was

identified in roasted coffee, where pronylated amino acids and proteins seemed to be present at very low amounts. Nmethylpridinium was shown to have strong

chemopreventive effects on modulating Phase II enzymes both in vitro and in vivo

(Somoza et al. , 2003).

1.2.3 Coffee – a source of MRPs

Coffee is one of the most popular beverages consumed in the world, and is known for its

desirable taste and aroma, stimulant effects and many potential health related benefits.

Recently, an increased number of papers have been published on various potential health benefits of coffee consumption (Tavani and La Vecchia, 2004; Ranheim and Halvorsen,

2005; van Dam and Hu, 2005; Cadden et al. , 2007; GomezRuiz et al. , 2008). Many

18 studies have shown that coffee consumption is associated with a reduced risk of several

chronic diseases (Giovannucci, 1998; Tavani and La Vecchia, 2004; van Dam and Hu,

2005; Barranco Quintana et al. , 2007; Cadden et al. , 2007; Larsson and Wolk, 2007).

Various physiological and pathological responses can be attributed to the bioactive compounds, including caffeine, chlorogenic acids (CGA), Maillard reaction products

(MRPs) and diterpenes kahweol and cafestol (K+C). Table 1.2 reports the overall chemical composition of green beans and roasted coffee beans (Arya and Rao, 2007).

Table 1.2 Composition of green and roasted coffee (adopted from (Arya and Rao, 2007)) Constituent Green (%DB) a Roasted (%DB) b Hemicellulose 23.0 24.0 Cellulose 12.7 13.2 Protein (nonalkaloid N) 11.6 3.1 Fat 11.4 11.9 Chlorogenic acids 7.6 3.5 Sucrose 7.3 0.3 Lignin 5.6 5.8 Caffeine 1.2 1.3 Trigonelline 1.1 0.7 Reducing sugars 0.7 0.5 Unknown 14.0 31.7 Total 100.0 100.0 a Dry Green Beans. b not corrected for dry weight roasting loss, which varies from 25%.

1.2.3.1 Composition of coffee bioactive components

1.2.3.1.1 Caffeine

Caffeine content in coffee beverages can be quite variable, depending on the type and source of coffee beans, roasting method and how the coffee is prepared (Barone and

Roberts, 1996; Harland, 2000; McCusker et al. , 2003). Robusta beans have a relatively

higher caffeine content than Arabica (Charrier, 1975). Dark roasted coffee contains less

19 caffeine than coffee made from light and medium roasted beans (Anon, 2004). The

caffeine content of coffee has shown to vary significantly between brands and the dayto day serving frequency (McCusker et al. , 2003). See Table 1.3.

Table 1.3 Caffeine content of different coffee beverages Coffee Caffeine (mg) in 8 oz serving Brewed 135 Groundroasted, percolated 118 Groundroasted, drip 180 Instant 106 Starbucks espresso 280 Starbucks, mocha, latte, Americano 35 Starbucks regular 130 Maxwell House regular 110 Big Bean regular 82 Data from (Barone and Roberts, 1996; Harland, 2000; McCusker et al. , 2003)

Caffeine is absorbed in the stomach and small intestine and metabolized primarily in the

liver (McCusker et al. , 2003). It is almost completely absorbed and distributed to all tissues, including the brain due to its relatively small size and optimal hydrophobic character. The plasma halflife of caffeine ranges from 2.3 to 12 h, depending on the physiological or health condition of the individual (Baselt, 2002). Peak caffeine plasma concentrations occur at 45 min to 2 h after ingestion (Ellenhorn and Barceloux, 1988).

Caffeine appears to exert most of the biological effects through the antagonism of the potent endogenous neuromodulator, adenosine (Dunwiddie and Masino, 2001). The effect of caffeine is generally stimulatory, including central nervous system stimulation, acute elevation of blood pressure, increased metabolic rate, and diuresis (Carrillo and

Benitez, 2000).

20 1.2.3.1.2 Chlorogenic acids (CGA)

Chlorogenic acids (CGA) are abundant phenolic compounds present in coffee, with

caffeoylquinic (CQA), feruloylquinic (FQA), and dicaffeoylquinic (diCQA) acids being the major subclasses. Depending on the coffee bean cultivar, green coffee beans contain between 614 % CGA on a dry matter basis (Farah and Donangelo, 2006). It has been

shown that the CGA concentration in Arabica green bean extracts varied between 16.6 %

to 22.4 % (w/w) due to different extraction procedures (with water) (Budryn et al. , 2009).

During roasting CGA undergoes a progressive destruction and transformation (Figure 1.1)

(George et al. , 2008). Nevertheless, coffee beverages are still a major dietary source of

CGA. In roasted Arabica coffee extracts, CGA ranged from 2.6 % to 15.8 % (w/w),

depending on the roasting degree and extraction methods (e.g. water) (Budryn et al. ,

2009). It has been estimated that one cup (240 ml, 8 oz) of groundroasted Arabica coffee

contains 80230 mg CGA, compared to 80400 mg in a cup (240 ml, 8 oz) of Robusta

coffee. Instant coffee can provide between 35110 mg CGA per gram of soluble powder

(Farah and Donangelo, 2006).

A recent study indicated that all major CGA in coffee are bioavailable and are absorbed

and/or metabolized differently in humans (Monteiro et al. , 2007). They reported that

there are two major temporal absorption patterns of CGA after coffee consumption,

which suggested an early absorption in the stomach followed by absorption throughout

the small intestine (Monteiro et al. , 2007). The C max (maximum plasma concentration) of total CGA was found to vary from 4.7 to 11.8 mol/L, among six individual human subjects and T max (time corresponding to C max ) for total CGA varied significantly between individuals (from 1 to 4 h) (Monteiro et al. , 2007). Chlorogenic acids were also

21 shown to be metabolized by the liver and gut microflora into various aromatic acid

metabolites (Gonthier et al. , 2003; Mateos et al. , 2006). Enterohepatic circulation of

CGA has been observed for up to 48 h after phenolic intake, suggesting that CGA

undergoes a gradual utilization and excretion in humans (Cremin et al. , 2001). The biological properties of dietary CGA will depend on the whole body kinetic flux of the phenolic acid which in turn involves absorption, metabolism, distribution and interaction

with target tissues (Cremin et al. , 2001; Monteiro et al. , 2007).

Chlorogenic acid

+ Quinic acid Caffeic acid Slow Degradation Rapid Degradation

+ + + +

Catechol 4-ethyi Catechol Hydro quinone Catechol Pyrogallol Gallic acid

Figure 1.1 CGA degradation during roasting of coffee bean (George et al. , 2008)

1.2.3.1.3 Maillard reaction prducts (MRPs) in coffee

During the roasting process of coffee, carbohydrates and protein are degraded and the

Maillard reaction that occurs leads to the formation of flavour and colored products

(Oosterveld et al. , 2003). These MRPs are responsible for the development of the

characteristic brown color and the basic taste of bitterness and astringency common to

22 coffee. The MRPs that are regarded as important contributors to the coffee flavour are the volatile aroma compounds (Yanagimoto et al. , 2002). During roasting, phenolics, especially CGA are partially degraded and bound to MRP polymer structures, thus contributing to some extent to the brown Maillard products (DelgadoAndrade et al. ,

2005; Bekedam et al. , 2008a; Bekedam et al. , 2008b). Studies have shown that melanoidins make up 25 % (w/w) of coffee dry matter (Borrelli et al. , 2002), and the concentration increases with increased roasting time (Sacchetti et al. , 2009). It has been shown that arabinogalactan is the most abundant sugar present in the melanoidinrich coffee fractions, and that the residual amount of this sugar is affected by roasting, with a consequent loss of arabinose (DeMaria et al. , 1996; Bekedam et al. , 2008a). Researchers

(Oosterveld et al. , 2003) showed that coffee polysaccharides are degraded during roasting and may be involved in MRP formation. In addition, studies have shown that amino acids present in coffee beans are degraded during roasting, and that the nitrogen from these amino acids may end up in melanoidin structures (Macrea, 1985; DeMaria et al. , 1996).

Arginine, lysine, serine, threonine, histidine and asparagine have also been shown to be significantly reduced during the roasting process of coffee beans (Macrea, 1985;

Bekedam et al. , 2006). Some researchers found that serine was the most affected amino acid during the roasting process, which also was suggested to be an important flavour precursor in coffee (DeMaria et al. , 1996). Coffee proteins, especially those that contain highly reactive εamino, thiol, or methylthiol groups, undergo chemical changes upon roasting and are likely to be involved in melanoidin formation (Rizzi, 1999; Bekedam et al. , 2006; Bekedam et al. , 2007).

23 Some products in the initial stage of the Maillard reaction (e.g. Amadori rearrangement product, ARP) are degraded via different pathways (Ames, 1992; Anese and Nicoli,

2003), thus, leading to the formation of reductones and furfurals. In vivo , these products are absorbed by diffusion and metabolized by the colonic microbiota (Erbersdobler and

Faist, 2001). Most metabolic transit data on melanoidins has been obtained in rats (Faist and Erbersdobler, 2001). Generally, melanoidins, of any source are characterized as having a low digestibility and bioavailability (Borrelli and Fogliano, 2005), with the relatively small fraction absorbed being speculated to be utilized to low degree since they are excreted in the urine in a slightly modified or unmodified form (Erbersdobler and

Faist, 2001).

1.2.3.1.4 Cafestol and kahweol (C+K)

Ditepene cafestol and kahweol (C+K) represent the major part of the unsaponifiable lipid fraction present in coffee beans. Commercial groundroasted coffees contain about 1%

(w/w) of diterpenes (Urgert et al. , 1995). These diterpenes comprise up to 1015% of the lipid fraction of roasted coffee beans (Lercker et al. , 1995). The brewing method is a major determinant of diterpene content in coffee beverages (Urgert et al. , 1995; Gross et al. , 1997). Diterpenes are extracted from ground coffee during brewing, but are mostly removed by paper filters. Turkish coffee, Boiled, and French press brews contain relatively high levels of C+K, while filtered, percolated, and instant coffee contain low levels of C+K (Urgert et al. , 1995; Gross et al. , 1997). Studies performed in ileostomy patients indicate that about 70% of the C+K in unfiltered coffee is absorbed intestinally

(De Roos et al. , 1998). Only a small part of the diterpenes is excreted in urine, which

24 indicated an extensive metabolism of C+K in human body (Ratnayake et al. , 1993;

Urgert et al. , 1996).

1.2.3.2 Coffee as a source of dietary antioxidant

1.2.3.2.1 Antioxidant intake in human diet

Coffee has been reported to have high antioxidant activity, which may be of great benefit in improving the quality of life of consumers by preventing, or postponing, the onset of many age related degenerative diseases. Researchers showed that coffee contained the greatest antioxidant potential among 34 common beverages (Pellegrini et al. , 2003). The antioxidant activity of coffee beverages was over six times greater than that of green tea, and about three times as high as that found in red wines. It is particularly noteworthy that coffee represents a major source of dietary antioxidant intake in Germany (Radtke et al. ,

1998), Spain (Pulido et al. , 2003), the United Kingdom (Clifford, 1999), and Norway

(Svilaas et al. , 2004) (Figure 1.2). Although the antioxidant properties of coffee have been attributed to caffeine, the formation of MRPs during roasting, and the relatively

great extent a number of phenolic compounds, likely supersedes the presence of caffeine

in terms of contributions to total antioxidant activity (Daglia et al. , 2000; del Castillo et al. , 2002; Caemmerer and Kroh, 2006).

25

A B

Figure 1.2 Contribution of coffee to the antioxidant intake in diet. A. Norway (Svilaas et al. , 2004); B. Spain (Pulido et al. , 2003).

1.2.3.2.2 Antioxidant property of phenolics in vitro

Chlorogenic acids are the predominant phenolics found in green coffee beans, which

contribute to most of the overall antioxidant activity (Caemmerer and Kroh, 2006). The

free CGA content in roasted coffee is lower than that in the green coffee bean due to the

degradation of CGA at thermal roasting temperatures. However, recent research has

demonstrated that the antioxidant activity of CGA was not completely destroyed despite

the chemical alterations that occur with heating (Bekedam et al. , 2008c). In other words,

CGA does not totally lose the phenolic nature, albeit, the active moiety that provides

antioxidant activity is not retained as free CGA. Instead, CGA degradation likely

involves a complex interaction where it is bound to other molecules via ionic and ester bonds (DelgadoAndrade et al. , 2005; Nunes and Coimbra, 2007; Bekedam et al. , 2008d).

Phenolic antioxidants therefore may still contribute to the overall antioxidant activity of

coffee beverages, but in a transformed state. In addition to the direct scavenging effect by

26 CGA on ROS and free radicals, which explain the affinity to inhibit the oxidation and peroxidation to lowdensity lipoprotein (LDL) (Castelluccio et al. , 1995) and decreased

ROSinduced DNA damage (Yamanaka et al. , 1997), other studies have shown that CGA

can upregulate some cellular xenobiotic phase II enzymes (Kitts and Wijewickreme,

1994; Feng et al. , 2005) and suppress ROS mediated NFκB, AP1, and mitogen

activated protein kinase (MAPK) activation (Feng et al. , 2005).

1.2.3.2.3 Coffee MRPs and the antioxidant potential

Many heterocyclic compounds derived from the Maillard reaction have been identified

and quantified in coffee brews, including pyrroles, oxazoles, furans, thiazoles, thiophenes,

imidazoles, and pyrazines (Fuster et al. , 2000). These compounds all possess antioxidant

activity, with pyrroles showing the highest activity relative to thiazoles and pyrazines

having the least activity in inhibiting hexanal oxidation (Fuster et al. , 2000; Yanagimoto

et al. , 2002). Underlying mechanisms for this apparent activity has been suggested to be

facilitated by the electron density of the carbon atoms present on the heterocyclic ring,

and different functional groups on the heterocyclic ring (Eiserich and Shibamoto, 1994;

Yanagimoto et al. , 2002). Under mild roasting conditions, CGA has been shown to be the

main component responsible for the free radical scavenging activity of coffee brews (del

Castillo et al. , 2002). However, MRPs may also be the principal component with free

radical scavenging activity in more severely roasted coffees (del Castillo et al. , 2005).

Some researchers (Borrelli et al. , 2002) found that the antiradical activity of coffee

melanoidins decreased as the intensity of roasting increased, but the affinity to prevent

linoleic acid peroxidation was higher in the dark roasted coffee samples. Pretreatment of

human HepG2 cells with digested coffee melanoidins prevented the increase in cell

27 damage evoked by tert butylhydroperoxide (Goya et al. , 2007). Researchers suggested that the antioxidant activity of coffee melanoidins could be attributed to the incorporated

CGA and CGA degradation products (DelgadoAndrade and Morales, 2005; Delgado

Andrade et al. , 2005). This incorporation may also enable CGA in the human colon to interact with gut microbiota, which plays an important role in maintaining health (Tuohy et al. , 2003). Moreover, new antioxidative structures formed through the Maillard reaction are also present in melaniodins (Nicoli et al. , 1997; Bekedam et al. , 2008c). Low molecular compounds released from coffee melanoidins after gastrointestinal digestion can exert antioxidant activity when assayed by five different methods, and the antioxidant activity was even higher than melanoidins and compounds ionically bound to melanoidins (RufianHenares and Morales, 2007b). Volatile coffee MRPs were proposed to possess potential antioxidant activity by preventing DNA damage in vitro

(Wijewickreme and Kitts, 1998c) and influence gene expression in rats brain (Seo et al. ,

2008).

1.2.3.2.4 Caffeine contributes to the antioxidant activity of coffee

Caffeine and metabolites exhibit both antioxidant and prooxidant properties in vivo , which depend on many parameters such as dose, level of atmospheric O 2 exposure, presence of transition metals, and the biological and chemical endpoints used for the

measurement of activity (George et al. , 2008). Caffeine is an effective inhibitor of lipid peroxidation as shown by the experiment where millimolar concentrations of caffeine

scavenged ROS (Devasagayam et al. , 1996). At physiological concentrations, caffeine

metabolites can prevent LDL oxidation (Lee, 2000). In general, the antioxidant ability of

caffeine was shown to be similar to that of the antioxidant glutathione (GSH), and

28 significantly higher than that of ascorbic acid (Devasagayam et al. , 1996). The

antioxidant activity of caffeine and its metabolites was probably attributed to the presence

of the carbonyl group at the C8 position of the pyridine ring (George et al. , 2008) (Figure

1.3). It is suggested that this chemical structure enables caffeine to scavenge highly

reactive free radicals, such as hydroxyl radical (OH•), and the generated caffeine radical

may be excreted in urine or stabilized by other antioxidants.

+

Figure 1.3 Mechanism of scavenging of free radicals by caffeine (George et al. , 2008).

1.2.3.3 Coffee consumption and health

To support many of the biological/biochemical activities attributed to coffee components,

it is noteworthy that many studies have shown that coffee consumption is associated with

reduced risk of several chronic diseases. Caffeine is the most widely studied coffee

component, however, it is not the major contributor to many beneficial health related

effects (Levin, 1982; Corrao et al. , 2001; Greer et al. , 2001). Table 1.4 summarizes the

quantitative assessments of the relationship between coffee consumption and the risk of

several diseases from a metaanalysis of epidemiologic studies. A systematic review of 9 prospective cohort studies, including more than 193,000 men and women, found that

habitual coffee consumption is associated with a substantially lower risk of Type 2

diabetes (van Dam and Hu, 2005). This association does not differ by sex, obesity, or

29 region (van Dam and Hu, 2005). Many casecontrol studies in Asia, Northern Europe,

Southern Europe, and North America have shown consistent inverse association between

coffee consumption and the risk of colorectal cancer, although the evidence from cohort studies is inconclusive (Giovannucci, 1998). A metaanalysis that combined the results of

12 casecontrol studies also found that frequent coffee consumers had a 28 % lower risk of colorectal cancer than infrequent coffee consumers (Giovannucci, 1998). Recently, a metaanalysis including 4 cohort and 5 case control studies found that an increase in

consumption of 2 cups of coffee per day was associated with a 43 % reduction in the risk of liver cancer (Larsson and Wolk, 2007). Coffee drinking has also been shown to have positive effect on neurodegenerative diseases, such as Alzheimer’s disease (Barranco

Quintana et al. , 2007). The available data suggest that this effect is due to caffeine intake

(Cunha, 2008; Arendash et al. , 2009). There is no data that has alluded to the presence of

coffee phenolics and MRPs contributing to these benefits.

In general, currently available evidence suggests that moderate amount of coffee

consumption has positive health benefits for most people. Caffeine is associated with

various aspects of mental health and brain function due to the effects on the central

nervous system. The presence of antioxidants such as CGA and MRPs may also be

important contributors for some of the health related effects attributed to coffee beverage

consumption.

30 Table 1.4 Summary of potential health benefits of coffee consumption from epidemiological studies

Health concerns Coffee consumption Relative risk Source Level of intake 1 (RR) (No. of studies) Type 2 diabetes Low * 1.00 Van Dam and Hu, 2005 Third highest * 0.94 (9 national cohort studies) Second highest * 0.72 Highest * 0.65 Colorectal cancer Low # 1.00 Giovannucci, 1998 High # 0.72 (12 casecontrol studies) Liver cancer Per 2 cups/day 0.57 Larsson and Wolk, 2007 increment $ (4 cohort and 5 casecontrol studies) Alzheimer’s > 0 cups/day Φ 0.73 Quintana et al ., 2007 disease (2 cohort and 2 casecontrol studies) 1 * The low level of consumption (reference) denotes 0 cups or 2 or less cups per day; the third highest level denotes 1 to 3 cups per day, or 3 or more cups per day, or 4 to 5 cups per day; the second highest level denotes 4 to 5 cups per day, or 5 to 6 cups per day; and the highest level denotes 6 or more, or 7 or more cups per day. # The low level (reference) denotes less than 1 cups per day; and the high level denotes 4 or more cups per day. $ The estimated RR is for an increment of 2 cups of coffee per day. Φ The RR is for coffee consumers (> 0 cups per day) versus nonconsumers.

31 1.3 RESEARCH HYPOTHESES AND OBJECTIVES

General thesis hypothesis

MRPs exhibit different antioxidant activities that can be attributed to differences in

molecular weight and chemical character. Coffee MRPs will modulate the antioxidant

status in Caco2 cells through the regulation of enzymatic antioxidants and genes that are

involved in oxidative stress and/or the antioxidant defense system.

General thesis objective

To characterize and compare the chemical properties of nonroasted coffee, roasted

coffee, model MRPs, and related ultrafiltration fractions. To assess the associated

chemical antioxidant activities and biological effects of these products in Caco2 cell

culture.

Experiment I. Chemical characteristics and antioxidant properties of coffee extracts

Hypothesis

1. The low molecular weight fractions (MW<1KDa) recovered from both light

roasted (LR) and dark roasted (DR) coffee beverages have greater in vitro

antioxidant potential in comparison with high molecular weight fractions

(MW>1KDa) derived from the same roasting conditions.

2. The low molecular weight components (MW<1KDa) in coffee bind non

covalently to the high molecular weight components (MW>1KDa) and thus

contribute to the in vitro antioxidant activity of the high molecular weight coffee

fractions.

32 3. Increasing the degree of roasting in coffee beverages will induce an increase in

specific antioxidant enzyme activities in cultured human Caco2 cells.

Objective

1. To determine the molecular weight distribution of LR and DR coffee beverages

using water and sodium chloride ultrafiltration systems.

2. To qualitatively define the chemical characteristics of coffee extracts that include

color development and quality, browning intensity, UV spectra, and fluorescent

spectra.

3. To evaluate the in vitro antioxidant activities of coffee extracts, and related

ultrafiltration fractions using ORAC, ABTS and reducing power assays.

4. To investigate the affinity of coffee extracts to modify cellular chemopreventive

antioxidant enzymes (e.g. SOD, CAT, GR, GPX) activities and GSH content.

Experiment II. Coffee constituents and modulation of oxidative status in Caco-2

cells

Hypothesis

1. MRPs are the major constituents in roasted coffee brew that contribute to the

antioxidant activity of coffee.

2. Low molecular weight coffee MRPs (MW<1KDa) have greater antioxidant

activity in comparison to high molecular weight coffee MRPs (MW>1KDa).

3. Coffee MRPs vary in the affinity to increase specific antioxidant enzyme

activities in cultured human Caco2 cells, and low molecular weight MRPs

33 (MW< 1KDa) have greater impact than high molecular weight MRPs

(MW>1KDa).

4. Coffee can modulate the oxidative status in Caco2 cells through the regulation of

genes involved in oxidative stress and/or antioxidant defense system.

Objective

1. To determine the molecular weight distribution of extracts derived from green

coffee beans, roasted coffee beans and model MRPs using water ultrafiltration.

2. To test the chemical characteristics of coffee and coffee model MRPs, including

color development, browning intensity, UV spectra, fluorescent spectra and the

presence of αdicarbonyl compounds.

3. To evaluate the in vitro antioxidant activities of green and roasted coffee

beverages, coffee model MRPs and related ultrafiltration fractions using ORAC,

ABTS and reducing power assays.

4. To investigate the cellular in vitro antioxidant activities of coffee beverages,

coffee model MRPs and related ultrafiltration fractions by examining potential

effects on SOD, CAT, GR, GPX activity and GSH content, and the protection

against reactive oxygen species induced oxidative stress in Caco2 cells.

5. To investigate the cellular reaction of coffee MRPs, particularly the influence on

the gene regulation of specific antioxidant enzymes and other genes involved in

oxidative stress and/or antioxidant defense system using Realtime PCR array.

34

CHAPTER II

CHEMICAL CHARACTERISTERCS AND ANTIOXIDANT PROPERTIES OF

COFFEE

35 2.1 Introduction

During the process of roasting coffee, green coffee beans are heated to 200250 ºC, for

0.7525 min, depending on the requirement for final roasting (e.g. light, medium, or dark).

Complicated physical and chemical changes take place in the roasted coffee beans, which include thermal degradation of natural phenolic antioxidants and generation of brown, flavored compounds, called Maillard reaction products (MRPs). In lieu of the antioxidant properties that exist for both phenolics and MRPs, the roasting process of coffee will result in a different final level of antioxidant activity. Previous studies have reported that the antioxidant activity of coffee is in fact mostly dependent on the roasting conditions, much more so than brewing methods and the sources of coffee beans (Sacchetti et al. ,

2009).

The effect of roasting on the antioxidant activity of coffee brew has been thoroughly investigated by many researchers but notwithstanding this, the results have been inconsistent. For example, some studies reported an increase in antioxidant activity of coffee brew with roasting (Borrelli et al. , 2002; SanchezGonzalez et al. , 2005), while others found that the antioxidant activity decreased with roasting (Richelle et al. , 2001;

Borrelli et al. , 2002). Some workers concluded that medium roasted coffee has the highest antioxidant activity (del Castillo et al. , 2002; Caemmerer and Kroh, 2006), while another study found that nonroasted green coffee beans possess higher antioxidant activity than the corresponding roasted samples (Daglia et al. , 2000). To complicated matters further, both green and roasted coffee beans contain a complex mixture of unknown number of chemicals, some of which may exhibit high antioxidant activity in one antioxidant test, while others exhibit a high antioxidant activity in a different test. In

36 general, MRPs derived from coffee have been demonstrated to exhibit primary

antioxidant activity towards metal prooxidant sequestering and direct free radical

scavenging activities (Wijewickreme and Kitts, 1998b; DelgadoAndrade et al. , 2005;

Takenaka et al. , 2005).

The development of color due to the generation of MRPs is an important result of the

roasting of coffee beans. The HunterLab color parameters (L, a, b, ∆E) has been used to

measure the color development of yellow to brown pigments from different stages of the

Maillard reaction, and the brown intensity is also widely used to monitor the development

of MRPs (MacDougall and Granov, 1998; Morales and van Boekel, 1998; Leong and

Wedzicha, 2000; Jing and Kitts, 2004a). Few studies have been carried out on

characterizing the development of fluorescent MRPs in coffee during roasting.

The aim of the present study was to characterize the chemical properties of two coffee brews that had undergone different roasting processes. This work attempted to correlate

the differences in browning of coffee beans with antioxidant capacity using three

different chemical based assays. Emphasis was on the distribution of MRP components based on molecular weight and the contribution towards the relative antioxidant activity

of the coffee brew. Finally, an extension of these studies was performed using a cell based procedure to determine if the chemical antioxidant activity measurements indeed corresponded to biological effects that could be attributed to changes in the enzymatic antioxidant defense system of human colon adenocarcinoma Caco2 cells. Caco2 cells were used in this thesis for evaluating the bioactivity of coffee constituents and the effects on gastrointestinal cells (Popovich and Kitts, 2004a, b).

37 2.2 Materials and methods

2.2.1 Coffee preparation

Roasted coffee beans ( Coffea Arabica ; roasted at designated “light”: 185 °C15 min and

“dark”: 210 °C15 min; reference: non official company disclosure), were purchased from a local store and ground to powder (fine grind) in a standard coffee grinder. A sample of coffee powder (55 mg) was extracted with 1.1 L hot water using No.4 cellulosetype coffee filters (Melitta, Canada). The fresh coffee extract obtained was rapidly cooled in an ice bath and centrifuged at 750 g for 45 min. Coffee brew supernatants (100 ml) were freezedried and the rest (1 L) was extracted with petroleum ether (3 × 300 ml) to remove the crude lipids. The defatted coffee extract was again freezedried and the yields of both crude and defatted extracts were determined gravimetrically. The recovered organic layer was concentrated to dryness using a rotary evaporator (Bϋchi Rotavapor R114, Bϋchi Labortechnik AG, Flawil, Switzerland) under vacuum at 40 ºC and the yield was recorded. Samples were stored at 4 °C until analysis was conducted.

2.2.2 Ultrafiltration

Freezedried defatted coffee extracts were dissolved in water, or in 2 M NaCl and fractionated by multiplestep ultrafiltration (Millipore, USA). The molecular weight cut off for each fraction was: 10KDa (YM 10), 1KDa (YM1), and 0.5KDa (YC 500), respectively. In a different experiment, NaCl was used to release the low molecular weight compounds ionically bound to the high molecular weight component.

Ultrafiltration separation was performed on the samples under a nitrogen pressure of 40

38 psi and individual fractions were collected and freezedried. The residues for molecular

weight fractions, Fraction I (MW>10KDa), Fraction II (1KDa

III (0.5KDa

°C until analysis.

2.2.3 Physical chemical analyses

2.2.3.1 Measurement of color

Color analyses on the ground coffee samples, crude and defatted coffee extracts were performed using a HunterLab Labscan 600 spectrocolorimeter (Hunter Associates

Lboratory Inc., Reston, Virginia). The instrument was calibrated with black and white

tiles. Color was expressed in L (L = 0 yields black and L = 100 indicates diffuse white), a

(negative values indicate green and positive values indicate red), b (negative values

indicates blue and positive values indicate yellow) Hunter scale parameters. The

colorimetric difference E was obtained through the equation: E = [(L)2 + (a)2 + (b)2]0.5 .

Five measurements were carried out on each sample.

2.2.3.2 Measurement of browning and UVvis spectra

Coffee brew samples were dissolved in distilled water at 0.5 mg/ml and 200 l of each

sample was placed into a 96well plate for the test. The UVvis absorbance over the range

of 250700 nm was recorded with a 5 nm interval. Indices of browning of the coffee

extracts and related fractions were determined using an absorbance maximum set at 420

nm (Multiskan Spectrum, ThermoLabsystem, Helsinki, Finland). A blank, containing

only distilled water was used to correct absorption readings.

39 2.2.3.3 Measurement of fluorescence

Coffee brew samples were dissolved in 3 ml of MilliQ water (0.25 mg/ml), to prevent

quenching effects. The solution was then measured at an excitation wavelength of 400 nm and emission wavelength range from 350 to 550 nm using a Shimadzu RF5301 spectrofluorophotometer (Kyoto, Japan). An average of three readings was recorded.

2.2.4 Chemical based antioxidant assays

2.2.4.1 Trolox equivalent antioxidant capacity (TEAC) assay

ABTS [2, 2´Azinobis(3ethylbenzothiazoneline6sulfonic acid)] radical cation

(ABTS•+) stock solution was prepared by mixing 5 ml of 7 mM ABTS (Sigma, St. Louis,

MO, USA) with 88 l of 140 mM potassium persulfate. This mixture was allowed to remain in the dark, at room temperature for 1224 h until the reaction was complete and the absorbance was stable. Fresh ABTS•+ working solution was prepared for each assay by mixing 600 l ABTS•+ stock solution with 40 ml distilled water to obtain an absorbance of at least 0.4 at 734 nm. The ABTS radical scavenging effect of coffee extracts at different concentrations (01.0 mg/ml) was calculated using the following equation:

% inhibition= (1 absorbance sample /absorbance control ) ×100

The standard curve was linear between 025 mM Trolox (SigmaAldrich, Oakville, ON,

Canada). Trolox equivalent antioxidant capacity (TEAC) =slope sample /slope control . Results

were expressed in mmol Trolox equivalent (TE)/g sample.

40 2.2.4.2 Oxygen radical absorbance capacity (ORAC) assay

ORAC assay measures the ability of antioxidant components in test materials that inhibit

the decline in fluorescence induced by a peroxyl radical generator, AAPH (2, 2azobis (2

amidinopropane) dihydrochloride) (Wako Chemcal Inc., Richmond, VA, USA). The

following reactants were added to each well in 96well blackwalled plates: 100 l

sample (final concentrations of 01.0 g/ml) in 75 mM phosphate buffer (pH 7.0) or

Trolox standard (final concentrations of 06.0 M) and 60ml fluorescein (Sigma, St.

Louis, MO, USA) (final concentration 60 nM). Each plate was incubated at 37 ºC for

15min; then 60 l AAPH (final concentrations 12 mM) were added and fluorescence

readings (Excitement wavelength = 485 nm, Emssion wavelength = 527 nm) were

continuously taken (060 min) using a fluorescence microplate reader (Huoroskan Ascent

FL, Labsystems). Data transformation and interception were performed according to the

method of Davalos et al. (2004) and ORAC values were expressed as mmol TE/g sample.

2.2.4.3 Reducing power (Gu’s RP assay)

Reducing power of the sample was tested using the method of Gu et al. (2009a) with some modification. Aliquots (e.g. 300 l) of samples over a concentration range of 02.0 mg/ml and standard chlorogenic acid solutions (CGA; Sigma, St. Louis, MO, USA) (0

1.0 mg/ml) were mixed with 300 l phosphate buffer (0.2 M, pH 6.6) and 300 l potassium ferricyanide (BDH, Product Code. B10204). The mixture was incubated at

50ºC for 20min. The reaction was terminated by adding trichloroethanoic acid (TCA;

Fisher, Nepean, ON, USA) solution (10 % w/v) and centrifuged at 3,000 rpm for 10 min.

60 l of the supernatant was mixed with 60 l distilled water and 12 l ferric chloride

(0.1% FeCl 3) in a 96well plate, incubated for 8 min in the dark, and the absorbance was

41 measured at 700 nm. Reducing power was expressed as g chlorogenic acid (CGA)/g

sample.

2.2.5 Cell based assays

2.2.5.1 Cell culture

Caco2 cells were obtained from ATCC (Manassas, VA) and maintained in Minimum

Essential Medium (MEM, Sigma, St. Louis, MO, USA), which were supplemented with

10 % fetal bovine serum (FBS, Gibco, Grand Island, NY. USA), penicillin (100 U) and

streptomycin (100 g/ml) (Gibco, Grand Island, NY. USA). Cells (passage 2440) were

cultured in an incubator (37ºC) under an atmosphere of 5 % CO 2 with 90 % humidity.

Culture media were changed every 48 h. For the measurement of cell viability, cells were

seeded in 96 well plates at a density of 5 × 10 5 cells/ml 24 h before coffee treatments. For enzyme activities and glutathione status, cells were seeded in a cell culture Petri dish at a density of approximately 1.0 × 10 5 cells/ml. Cells were allowed to reach 80 % confluence, which took about 5 days in culture at the time of treatment. At subconfluence, cells were in homogeneously undifferentiated (Vachon and Beaulieu, 1992).

2.2.5.2 Cell counting

Culture medium was removed and 0.25 % TrypsinEDTA (Gibco, Grand Island, NY.

USA) was added to cells. Cells were incubated at 37 ºC until detachment from the plates occurred. Cells were manually dispersed to attain a single cell suspension. The trypsin was neutralized by adding fresh culture media to cells, which was followed by cell counting using a haemocytometer. Viable cells were assessed by trypan blue exclusion dye (Sigma, UK.).

42 2.2.5.3 Coffee treatment of Caco2 cells

Caco2 Cells were exposed to filter sterilized defatted coffee extracts (0.1 mg/ml in culture medium) at different incubation times. The treated cells were rinsed with icecold

PBS and then scraped into a 1.5 ml tube. The cell samples were put through a freeze

(liquid nitrogen, 2 min)thaw (37 ºC, 5 min) cycle, 3 times, to release cytosol constituents and then centrifuged at 4ºC at 15,000 g for 10 min. The supernatant was decanted into a new tube, adjusted to a final volume of 700 l, and the cell pellet was discarded. The supernatants were kept on ice prior to protein content, enzyme activity and glutathione status measurements.

2.2.5.4 MTT cell viability assay

Following incubation of cells with coffee extracts at different concentrations for different incubation periods, Caco2 cells were rinsed with phosphate buffer saline (PBS, pH 7.2).

A medium containing MTT (0.5 mg/ml; [3(4,5dimethylthiazol2yl)2,5diphenyl tetrazolium bromide], Sigma, St. Louis, MO, USA) was added to the cell culture. Cells were incubated in the dark for 4 h with the MTT medium. To solubilize the formazan crystal, SDS (10 %) in hydrochloric acid (HCl) (0.1 N) was added and the plates were incubated overnight. Optical density readings were taken at 570 nm in a microplate reader (ThermoLabsystems Multiscan Spectrum, Thermolabsystem, Chantilly, VA).

Absorbance values measured at 570 nm were corrected for background absorbance using well that containing only MTT medium. Cell MTT response (% control) was calculated from the equation:

% control = absorbance treatment /absorbance control × 100%

43 2.2.5.5 Protein content

Protein content of the Caco2 cell supernatants was measured according to the method of

Bradford using bovine serum albumin (BSA) as the standard protein (Bradford, 1976).

Briefly, 5 l of standard, sample or blank was mixed with 250 l Bradford dye reagent.

After 45 min incubation, the absorbance was read in a 96well plate in triplicate using a microplate reader at 595 nm. The standard curve was prepared in a range between 0.1 and

1.4 mg/ml BSA in PBS.

2.2.5.6 Glutathione status

Glutathione (GSH) status was measured by determining 5thio2 nitrobenzoate (TNB)

generation, resulting from the reaction of 5, 5’dithiobis (2nitrobenzoic acid) with GSH

(Anderson, 1985). Coffee treated and untreated cell supernatants were first deproteinized with 5% 5sulfosalicylic acid (SSA) solution (2:1 v/v), centrifuged to remove the precipitated protein and then assayed for GSH using the enzymatic procedure. In a 96 well plate, 10 l of supernatant was added to 240 l of a freshly prepared reaction

mixture, containing 30 l of 6 mM 5, 5’dithiobis (2nitrobenzoic acid) (DTNB) (Sigma,

St. Louis, MO, USA), and 210 l working solution [(0.248 mg/ml βnicotinamide adenine dinucleotide phosphate (NADPH; Sigma, St. Louis, MO, USA) in a sodium phosphate buffer (143 mM, pH 7.5)] and incubated at 37 ºC for 20 min. GSSH reductase

(0.5 U; Sigma, St. Louis, MO, USA), was added to initiate the assay. The reaction mixture, without sample, was used as a blank. TNB formation rate of both samples and

GSH standards (Sigma, St. Louis, MO, USA), were monitored using a microplate reader at 412 nm at 1 min intervals for 5 min. Cellular GSH concentrations (nmol/mg protein) were determined from a standard curve of nanomoles of GSH equivalents versus rate of

44 change in activity (e.g. change in absorbance/min). The calculations are shown below.

Means and ranges for SD were obtained from two or more independent experiments.

Each experiment was performed in duplicate.

nmoles GSH per mg of sample = A 412 /min (sample) × df A 412 /min (1 nmole) × pc

Where A 412 /min (sample) = slope generated by sample

A 412 /min (1 nmole) = slope calculated from standard curve for 1 nmole of GSH

df = dilution factor of original sample

pc= protein content of sample in the reaction in mg

2.2.5.7 AntioxidanteEnzyme assays

Catalase (CAT) activity was determined using the UV spectrophotometric assay reported by (Claiborne, 1985) with some modifications. A volume of 4.5 l of 50 % ethanol was added to 225 l of cell supernatant and incubated on ice for 30 min. This was followed by the addition of 10 % Triton X100 to a final concentration of 1.0 %, and mixed thoroughly. To initiate the reaction, a working solution (275 l) of 20 mM hydrogen peroxide (Fisher, Nepean, ON) in 50 mM phosphate buffer (pH 7.0) was added to individual wells on a 96well UV plate, followed by 125 l of the 10 times diluted cell supernatant. Absorbance readings were monitored at 240 nm at 25 °C for 2 min with a reference containing only the working solution. Catalase activity in samples was expressed as micromoles of hydrogen peroxide consumed per minute per milligram of protein.

Superoxide dismutase (SOD) activity was measured using a SOD Assay KitWST obtained from Dojindo (Gaithersburg, MD, USA). Generally, 100 l of

45 chloroform/ethanol (3/5, v/v) was added to 250 l of cell supernatant with vigorous mixing. The mixture was centrifuged at 1400 g at 4 ºC for 10 min. The supernatant was transferred into a new tube and the pellet was discarded. An aliquot of supernatant (e.g.

20 l) from different dilutions was incubated at 37 °C for 20 min with a reaction mixture containing , , and (2(4Iodophenyl) 3(4nitrophenyl)5(2,4 disulfophenyl)2Htetrazolium, monosodium salt) (WST1). Absorbance readings at 450 nm were recorded using a microplate reader. SOD derived from bovine erythrocytes

(Sigma, St. Louis, MO, USA) was used to make the standard curve. SOD activities of the samples were calculated based on the SOD standard. The definition of SOD Unit, according to Sigma product information, is that one unit inhibits the rate of reduction of cytochrome c by 50 % in a coupled system, using xanthine and xanthine oxidase.

The glutathione peroxidase (GPX) activity assay was performed by adding cell supernatant samples into a working solution that included 1 mM glutathione, 2 mM sodium azide (Fisher, Nepean, ON), 1 U/mL glutathione reductase, 0.1 mM β nicotinamide adenine dinucleotide phosphate (NADPH). The reaction medium was incubated for 5 min at 37 ºC before initiating the reaction with the addition of 10 l of 7.5

M hydrogen peroxide. The disappearance of NADPH absorbance at 340 nm was monitored for 3 min (Paglia and Valentine, 1967). A nonenzymatic reaction rate (blank) was determined by substituting water for cell supernatant and recording the decrease in

NADPH absorbance. One unit of GPX activity was defined to be equivalent to the oxidation of 1 nanomole NADPH per minute per milligram protein.

The glutathione reductase (GR) activity assay was determined using a Glutathione

Reductase assay kit obtained from Sigma (St. Louis, MO, USA). This assay is based on

46 the reduction of glutathione (GSSG) by NADPH in the presence of glutathione reductase.

In addition, DTNB reacts spontaneously with the reduced glutathione (GSH) and generates 5thio (2nitrobenzoic acid) (TNB), which can be measured by the increase in absorbance at 412 nm using an extinction coefficient (ε mM ) of 14.15 for TNB. One unit causes the reduction of 1.0 mole of DTNB to TNB at 37 °C.

2.2.6 Statistical analysis

Each experiment was performed in triplicate (e.g. three wells or three cell culture plates) and repeated three times in separated experiments. Collected data were expressed as mean + SD. Data were analyzed by Oneway Analysis of Variance (ANOVA), followed by Tukey’s pairwise comparisons. Comparison of multiple treatments with control was

done by oneway ANOVA, followed by Dunnett’ test. Significant differences between

two samples were analyzed with Student ttest. In some cases, data were analyzed with

twoway ANOVA followed by Bonferroni posttests. The level of confidence required

for significance was selected at p<0.05.

47 2. 3 Results

2. 3.1 Yields and recovery of coffee brews and ultrafiltration fractions

Freshly brewed coffee processed at specific temperature and time conditions required to produce a light roast (LR) and dark roast (DR), respectively, resulted in a final yield of crude coffee extracts that was 13.48 g/L and 14.68 g/L, respectively (Table 2.1). The weight of total crude lipids recovered from coffee brews defatted by triple extraction with petroleum ether was not significantly different between LR and DR coffee extracts.

Similar yields of hydrophilic constituents were also obtained in LR and DR coffee extracts.

Water ultrafiltration was used to recover four distinct fractions of the hydrophilic coffee extracts with molecular weights that ranged from >10KDa (Fraction I); 110KDa

(Fraction II); 0.51KDa (Fraction III) and <0.5KDa (Fraction IV). The recovery of freezedried fractions is presented in Table 2.2. A relatively greater proportion (p<0.05) of the coffee brew constituents was found in the low molecular weight fractions (e.g.

Fractions III and IV). Total recovery of all coffee brew constituents in the four molecular weight fractions was 97 % and 96 %, for LR and DR, respectively. In order to remove noncovalent bound low molecular weight compounds from the melanoidin skeleton, the defatted coffee extracts were also treated with 2 M NaCl, followed by similar multistep ultrafiltration. The recovery of all coffee constituents present in the salt Fraction I S to IV S

(I S: S = salt) are presented in Table 2.2. Fraction IVS was not able to be recovered due to the excess salt content. A correction, therefore that consisted of summing the recovery of the three fractions and expressing this as 100 percent recovery was used to calculate and estimate Fraction IV S recovery. The recovery from the salted Fraction IVS was found to

48 be numerically greater (p<0.05) than that recovered in the nonsalted Fraction IV using water ultrafitration. Thus, a greater recovery of low molecular weight compounds

(MW<0.5KDa) was made when coffee brew extracts were treated with salt to liberate noncovalently bound compounds.

Table 2.1 Recovery yields of coffee extracts 1 Coffee extracts Yield 2 LR DR Crude extract 13.48+0.77 b 14.68+0.28 b Defatted extract 12.96+0.38 b 13.87+0.12 b Total crude lipid 0.37+0.16 a 0.67+0.09 a 1 Values are expressed as mean + SD, n=3. ab represent means in columns that are significantly different. Statistical analyses were done by twoway ANOVA with Bonferroni posttests; level of confidence set as 0.05. 2 Yields of crude, defatted extracts and total crude lipids are expressed as g/L of coffee brew.

Table 2.2 Recovery of coffee fractions by water and salt ultrafiltrations 1 Fractions (MW) 2 LR DR water salt water salt I (>10KDa) 17.7+0.6 a 15.5+5.9 b 19.4+1.5 ab 18.0+4.6 a II (110KDa) 13.8+2.2 a 8.0+2.5 a 15.6+2.7 a 12.5+1.9 a III (0.51KDa) 30.7+0.9 b 28.6+4.5 c 21.5+2.2 b 18.8+2.5 a IV (<0.5 KDa) 3 34.9+5.4 bx 48.0+1.1 dy 39.8+2.3 cx 50.7+3.7 by

1 The recovery of fractions are expressed as % of defatted extracts dry weight (mean + SD, n=3). ab represent means in columns that are significantly different; xy represent different means in rows between water and salt ultrafiltration. Statistical anaylyses were done by twoway ANOVA with Bonferroni posttests; level of confidence set as 0.05. 2 Fractions were derived from defatted extracts, based on molecular weight (MW). 3 The recovery of Fraction IV derived from salt ultrafiltration system was calculated using this equation: 100 – R Fraction I – R Fraction II – R Fraction III ; R = recovery.

49 2.3.2 Chemical characteristics of coffee

2.3.2.1 Colorimetric measurements (Lab analysis)

Quantitative colorimetric measurements of LR and DR ground coffee powder, and

freezedried crude and defatted extracts and related ultrafiltration fractions were obtained

to assess differences in color. The measure of total color difference (E), reveals all

colorimetric parameters such as chroma and lightness (L). Significantly higher L and E

values were observed in the ground powder and freezedried crude extracts derived from

LR coffee beans compared to those from DR beans (Table 2.3). However, no significant

differences in L and E values were observed between LR and DR coffee defatted

extracts. Colorimetric parameters of water ultrafiltration fractions derived from defatted

coffee extracts are present in Table 2.4. It was found that a pattern of relative increase in both L and E corresponded to a decreased molecular weight in the four fractions.

Table 2.3 Color parameters (L, E) of coffee ground powder and extracts 1 Coffee extracts L E LR DR LR DR Ground power 16.2+1.0 ay 11.3+0.5 ax 21.3+1.0 ay 13.3+0.6 ax Crude extract 35.5+2.3 cy 30.5+1.2 bx 43.0+3.0 cy 37.8+1.6 bx Defatted extract 31.7+0.7 b 29.5+2.2 b 38.7+0.7 b 35.8+2.7 b

1 Values are expressed as mean + SD, n=3. abc represent means in columns that are significantly different; xy represent significantly different means in rows between LR and DR. Statistical analyses were done by twoway ANOVA with Bonferroni posttests; level of confidence set as 0.05. Color difference is expressed as E = (L 2 + a 2 +b 2)1/2 , where “L” represents lightness, “a” represents red/green, and “b” represents yellow/blue.

2.3.2.2 Browning of coffee extracts (UV analysis)

The extent of browning in LR and DR coffee ultrafiltration fractions is shown in Table

2.4. There was no significant difference observed for UV absorbance between defatted

50 LR and DR coffee extracts. The degree of browning found in Fractions I and II was

significantly higher (p<0.05) than that of fractions III and IV. No significant difference in

browning was found between Fractions I and II, and between low molecular weight

Fractions III and IV.

Table 2.4 Color parameters (L, E) and browning of fractionated coffee extracts 1 Fractions(MW) 2 L E 3 Browning 4 LR DR LR DR LR DR I (>10KDa) 29.5+3.1 a 23.0+0.8 a 34.7+3.5 ay 27.1+0.9 ax 0.54+0.07 b 0.54+0.07 b II (110KDa) 35.4+1.4 ab 32.7+2.7 b 43.1+1.8 b 39.3+2.5 b 0.50+0.09 b 0.39+0.08 b III (0.51KDa) 37.3+2.9 bc 32.9+5.5 b 52.3+3.9 cy 40.2+3.9 bx 0.13+0.02 a 0.14+0.00 a IV (<0.5 KDa) 42.5+3.7 c 46.0+5.2 c 56.2+3.0 c 60.7+3.2 c 0.07+0.00 a 0.07+0.00 a 1 LR and DR represent light roasted and dark roasted coffee. Values represent mean + SD, n = 3. abcd The means with different superscript letters in each column are significantly different; xy represent significantly different means in rows between LR and DR (p<0.05, twoway ANOVA with Bonferroni posttests). 2 Fractions are derived from defatted extracts, based on molecular weight (MW). 3 Color difference is expressed as E = (L 2 + a 2 +b 2)1/2 , where “L” represents lightness, “a” represents red/green, and “b” represents yellow/blue. 4 Browning intensities are absorbance readings at 420 nm. Samples were diluted to 0.5 mg/ml with distilled deionized water prior to spectrum measurement.

2.3.2.3 Fluoresence and UVvis spectra

The defatted extracts and fractions of different molecular weight components derived

from coffee brews produced similar spectral fluorescence patterns with a sharp peak

occurring at 400 nm and another broad peak also observed at higher emission

wavelengths (Figure 2.1). The intensity of the first peak at 400 nm in Fraction I was

greater than that found for other fractions. In general, the fluorescence intensity for

different coffee fractions did not correspond to the trend found between increased

molecular weight and the degree of browning and tristimulus colorimetry for Fractions I

IV.

51 The UV absorbance spectra measured over a 250700 nm range was obtained for both LR

and DR coffee defatted extracts and ultrafiltration fractions (Figure 2.2). There was no peak absorption associated with Fraction I for both LR and DR coffee extracts. Fraction

II had a characteristically higher absorbance intensity but a similar pattern existed between LR and DR coffee extracts. In contrast, the absorbance spectra for Fraction III contained distinct absorbance maxima at 285 nm and 320 nm for LR coffee; while the

DR coffee had an absorbance maximum at 280 nm (Figure 2.2D). A clear difference in the intensity of absorbance spectra were found in Fraction IV between LR and DR.

Maximal absorbance values recorded at 275 nm were greater for LR than DR (p<0.05).

52

Figure 2.1 Fluorescence emission spectra (350550 nm) of light roasted (LR ) and dark roasted (DR ) coffee extracts and fractions. Defatted coffee extracts (A), Fraction I (B), Fraction II (C), Fraction III (D) and Fraction IV(E), measured with the excitation wavelength set at 400 nm. Samples were diluted to 0.25 mg/ml with miliQ water prior to spectrum measurement.

53

Figure 2.2 Comparison of the UVvisible spectra of light roasted (LR ) and dark roasted (DR ) coffee extracts and fractions. Defatted coffee extracts (A), Fraction I (B), Fraction II (C), Fraction III (D) and Fraction IV(E), measured from 250700 nm absorbance wavelength. Samples were diluted to 0.5 mg/ml with distilled deionized water prior to spectrum measurement.

54 2.3.3 Antioxidant activity of coffee extracts in chemical systems

Tables 2.42.6 present the results of antioxidant activity measurements on coffee samples

using three selected methodologies. ORAC and TEAC data are expressed as mmol

equivalents of Trolox (TE)/ g of sample, a commonly used unit for antioxidant assays.

Chlorogenic acid (CGA) equivalent was used for the RP assay, due to the presence of this phenolic in coffee and relative important contribution to the overall antioxidant capacity.

The antioxidant activity of LR and DR coffee was compared to see the influence of

roasting conditions. Generally, LR coffee had greater (p<0.05) antioxidant activity

compared to DR, based on the results from the three assays (Table 2.4). There was no

significant difference in antioxidant activity between crude and defatted coffee brews, as

measured by the three assays (Table 2.4), thus indicating that the hydrophilic components

in coffee brew were mainly responsible for the observed antioxidant activity. However,

significant differences in the antioxidant activity were found between different molecular

weight fractions. For the ORAC assay, Fractions III and IV had higher TE values in

comparison to the corresponding Fractions I and II (p<0.05), thus indicating a higher peroxyl radical scavenging ability associated with low molecular weight constituents

(MW<1KDa). For the TEAC assay, Fraction I exhibited the lowest ABTS radical scavenging capacity (p<0.05), and Fractions II and III had the highest TEAC values

(p<0.05). Fraction I had the lowest reducing activity. Fraction III had the highest antioxidant capacity based on all antioxidant measurements. In order to evaluate the efficiency of ultrafiltration at maintaining the antioxidant activity of different coffee components present in all fractions, Fractions I, II, III and IV were combined in proportion to form a recombined fraction. No significant difference in the antioxidant

55 activity between the recombined fraction and the nonfractionated defatted extract was obtained in all three assays (Table 2.5). This result indicated that the antioxidant activity of coffee components was maintained after ultrafiltration.

Samples that were treated with 2 M NaCl to release low molecular weight compounds

ionically attached to the melanoidin skeleton were also assayed for antioxidant activities.

In the DR coffee extracts, TEAC and RP, but not ORAC, increased significantly (p<0.05)

in Fraction I S after salt treatment. In contrast, Fraction I S recovered from the LR coffee extract showed increased ORAC and TEAC values (p<0.05) compared to the nonsalt treated fraction I. For Fraction II S, both TEAC and RP values were significantly higher

(p<0.05) in LR and DR extracts compared to nonsalt treated samples. The ORAC value

taken for Fraction III S of DR coffee extract was less than nonsalt treated Fraction III, while the TEAC and RP values were significantly higher (p<0.05). For LR Fraction III S, all the three values were significantly lower compared to nonsalt treated samples

(p<0.05).

56 Table 2.4 Antioxidant activities of coffee extracts and fractions 1 Coffee extract Assays 2 ORAC TEAC RP LR DR LR DR LR DR Crude 1.61+0.08 y 1.32+0.09 x 0.54+0.02 y 0.47+0.04 x 0.21+0.01 y 0.13+0.02 x Defatted 1.61+0.03 y 1.37+0.09 x 0.60+0.01 y 0.47+0.02 x 0.21+0.00 y 0.15+0.01 x Fractions (MW) I (>10KDa) 0.51+0.04 a 0.64+0.03 a 0.38+0.02 a 0.37+0.01 a 0.14+0.01 ay 0.11+0.00 ax II (110KDa) 1.25+0.06 b 1.31+0.06 b 0.62+0.04 cy 0.56+0.03 cx 0.17+0.01 b 0.15+0.01 b III (0.51KDa) 2.02+0.12 cy 1.82+0.09 cx 0.59+0.01 c 0.55+0.02 c 0.25+0.01 dy 0.15+0.01 bx IV (<0.5 KDa) 1.89+0.10 c 1.88+0.11 c 0.51+0.02 b 0.46+0.03 b 0.20+0.01 cy 0.14+0.01 bx 1 DR and LR represent dark roasted and light roasted coffee. Value represents mean + SD (n=3), significant differences were analyzed with twoway ANOVA with Bonferroni posttests. abc represent means in columns that are significantly different; xy represent significantly different means in rows between light roasted (LR) and dark roasted (DR). 2 ORAC represents oxygen radical absorption capacity; TEAC represents Trolox equivalent antioxidant capacity; RP represents reducing power. ORAC and TEAC values are expressed as mmol Trolox equivalents/g freeze dried coffee samples; RP values are expressed as mg CGA/g freeze dried coffee samples.

57 Table 2.5 Antioxidant activity of defatted nonfractionated coffee extracts and recombined extracts 1 Coffee extract Assays 2 ORAC TEAC RP LR DR LR DR LR DR Nonfractionated 1.61+0.03 y 1.37+0.09 x 0.60+0.01 y 0.47+0.02 x 0.21+0.00 y 0.15+0.01 x Recombined 3 1.64+0.14 y 1.39+0.10 x 0.61+0.04 y 0.46+0.03 x 0.19+0.01 y 0.14+0.01 x 1 Value represents mean + SD (n=3), significant differences were analyzed with twoway ANOVA with Bonferroni posttests. xy represent significantly different means in rows between LR and DR. No significant differences between defatted and fraction IIV. 2 ORAC represents oxygen radical absorption capacity; TEAC represents Trolox equivalent antioxidant capacity; RP represents reducing power. ORAC and TEAC values are expressed as mmol Trolox equivalents/g freeze dried coffee samples; RP values are expressed as mg CGA/g freeze dried coffee samples. 3 Recombined extract represents recombined fraction I, II, III and IV.

58 Table 2.6 Antioxidant activities of coffee fractions by water and salt ultrafiltration 1 Fractions 2 Assays 3 ORAC TEAC RP LR DR LR DR LR DR I 0.51+0.04 a 0.64+0.03 0.38+0.02 a 0.37+0.01 a 0.14+0.01 y 0.11+0.00 ax S by x b b by I 0.97+0.06 0.66+0.13 0.54+0.04 0.56+0.04 0.15+0.01 x 0.17+0.01 II 1.25+0.06 a 1.31+0.06 0.62+0.04 a 0.56+0.03 a 0.17+0.01 ay 0.15+0.01 ax S by x b b bx II 1.58+0.15 1.17+0.09 0.77+0.07 0.68+0.03 0.23+0.01 by 0.21+0.02 III 2.02+0.12 by 1.82+0.09 bx 0.59+0.01 by 0.55+0.02 ax 0.25+0.01 by 0.15+0.01 ax S III 1.23+0.05 a 1.08+0.08 a 0.51+0.01 ax 0.66+0.01 by 0.13+0.00 ax 0.21+0.00 by 1 Value represents mean + SD (n=3), significant differences were analyzed with twoway ANOVA with Bonferroni posttests. ab represent means in columns between fraction and fraction S that are significantly different; xy represent means in rows between light roasted (LR) and dark roasted (DR) that are significantly different. 2 Fractions were derived from defatted coffee extracts using water (I, II, III) or salt (I S, II S, III S), based on molecular weight (MW). 3 ORAC represents oxygen radical absorption capacity; TEAC represents Trolox equivalent antioxidant capacity; RP represents reducing power. ORAC and TEAC values are expressed as mmol Trolox equivalents/g freeze dried coffee samples; RP values are expressed as mg CGA/g freeze dried coffee samples.

59 2.3.4 Biological effects of coffee extracts

2.3.4.1 MTT response

The potential cytotoxicity of coffee was tested using defatted coffee extracts on Caco2 cells. The Caco2 cell MTT response to coffee extracts (LR and DR) was found to be dependent on both the incubation time and sample concentration (Figure 2.3). Cells showed reduced (p<0.05) viability when incubated with 2.5 mg/ml coffee (LR and DR) for 3 h. This toxicity potential was much more severe when cells were exposed to coffee extracts for longer times. For example, after 72 h treatment, cells showed a reduced viability at concentrations as low as 0.25 mg/ml coffee treatment (p<0.05). There were no signs of a toxic effect when cells were treated with coffee extracts at concentrations lower than 0.1 mg/ml for all incubation time. The IC 50 for different treatment durations of LR and DR coffee extracts, derived from the concentration response MTT curves, is summarized in Table 2.7.

60

Figure 2.3 Effects of light roasted (LR) and dark roasted (DR) coffee extracts on the tetrazolium reduction rate in the MTT assay after 3 h ( ), 12 h ( ), 24 h ( ), 48 h ( ) and 72 h ( ) incubation. A: LR; B: DR. Data (% control) represent means + SD obtained for triplicate samples measured in triplicate.

61 1 Table 2.7 IC 50 of coffee extracts on Caco2 cells using MTT assay

Time IC 50 (mg/ml) LR DR 3h 5.76+0.33 d 5.29+0.32 d 12h 2.94+0.20 c 3.12+0.15 c 24h 1.58+0.09 b 1.65+0.09 b 48h 1.27+0.07 a 1.24+0.06 a 72h 0.96+0.06 a 0.99+0.05 a

1 IC 50 is determined as the concentration of coffee that was required to reduce cell viability to half that of control in the MTT assay. Values represent means + SD obtained for triplicate samples measured in triplicate. abcd represent means in columns that are significantly different. No significant difference between light roasted (LR) and dark roasted (DR). Statistical analyses were done by twoway ANOVA with Bonferroni post tests; level of confidence set at 0.05.

2.3.4.2 Antioxidant enzyme activity

There was no significant change in the cellular antioxidant enzyme activities of GPX, GP,

and SOD (data in Appendix Table 13), following exposure to 0.1 mg/ml LR and DR

coffee at different incubation time periods. The activity of CAT was not changed

significantly in Caco2 cells following 3 h and 24 h coffee treatment, however, it was

decreased significantly (p<0.05) after 72 h of LR coffee exposure (Figure 2.4). A similar

result was not obtained with cells exposed to the DR coffee extract.

62

Figure 2.4 Effect of coffee extracts on catalase (CAT) activity in Caco2 cells. Cells were exposed to 0.1 mg/ml light roasted (LR) and dark roasted (DR) coffee for 3 h (A), 24 h (B) and 72 h (C). Cells incubated in culture medium were used as control. Results are mean + SD, n=3. Statistical analysis was done by oneway ANOVA with Dunnett’s multiple comparison tests. * indicates significant difference compared to corresponding control (p<0.05).

63 2.3.4.3 Intracellular antioxidant status

As an index of the intracellular antioxidant defense, the concentration of total glutathione

(GSH) was measured in Caco2 cells treated with 0.1 mg/ml LR and DR coffee extracts

for different time periods (Figure 2.5). No significant change in cellular GSH content was found after 3 h exposure to LR and DR coffee treatments, respectively. A significant

decrease (p<0.05) in GSH content was observed in cells treated with LR coffee for 24 h.

DR coffee treated cells showed a trend for a lower GSH content, but this effect was not significant. After 72 h exposure of cells to LR and DR coffee defatted extracts, a trend for reduced cellular GSH content was observed; however, the effect was again not significant.

64

Figure 2.5 Effect of coffee extracts on glutathione (GSH) content in Caco2 cells. Cell were exposed to 0.1 mg/ml light roasted (LR) and dark roasted (DR) coffee for 3 h (A), 24 h (B) and 72 h (C). Cells incubated in culture medium were used as control. Results are mean + SD, n=3. Statistical analysis was done by oneway ANOVA with Dunnett’s multiple comparison tests. * indicates significant difference compared to corresponding control (p<0.05).

65 2.4 Discussion

2.4.1 Chemical characteristics of coffee

Color change is an important phenomenon during the roasting process of green coffee beans. This is mostly due to the Maillard browning reaction (Massini et al. , 1990).

Colorimetric measurement is a simple and widely used technique to describe the roasting degree of coffee, which is classified as light, medium and dark according to a lightness parameter L. During the roasting process of coffee beans, lightness parameter L has been found to be negatively correlated to the roasting time at a given temperature and a linear correlation between the reciprocal of lightness and roasting time has also been reported

(Gokmen and Senyuva, 2006; Sacchetti et al. , 2009). In the present study, greater

lightness and total color (E) were associated with LR coffee beans that were roasted at a

lower temperature compared to DR beans. In addition, this result shows that freezedried

coffee extracts, which contain mainly water soluble MRPs, have greater lightness and

total color compared to ground coffee powder, which contain both water soluble and

insoluble MRPs. The color differences (L and E) between LR and DR coffee extracts

were not significant after delipidation, indicating some contributions of the lipophilic

coffee components to the color appearance of coffee extracts. These results also indicate

that the color parameters (L and E) of watersoluble MRPs derived from LR coffee

extract are not significantly different from DR coffee MRPs when roasted under similar

roasting conditions with the only exception of temperature. The present work suggests

that both the color parameters (L and E) and the degree of browning of coffee MRPs are

related to complex mixtures of MRPs components and their molecular weights. High

molecular weight MRPs had a greater degree of darkness and browning, while greater

66 lightness and less browning was associated with low molecular weight MRPs. The low molecular weight components of browning have been demonstrated to be generated in the

early stage of the Maillard reaction (Hayashi and Namiki, 1986).

Both fluorescence and UV absorbance have been used to characterize MRPs (Pongor et

al. , 1984; Wijewickreme et al. , 1997). In the case of coffee, at least two groups of compounds with different chemical structures showed typical patterns in the fluorescent spectra. One group that prevailed in the high molecular weight (MW>10KDa) fraction, and to a lesser extent in the other three fractions (MW<10KDa) had a sharp peak at 400 nm wavelength. The broad peaks that occurred at higher emission wavelengths (e.g. 450

500 nm) indicated that many fluorescent constituents were present in Fractions II, III and

IV derived from coffee extracts. Similar fluorescence spectra have been shown for different sugarLys model MRPs (Jing and Kitts, 2004a). The UV absorbance spectra of the high molecular weight (MW>1KDa) fractions derived from both LR and DR coffee extracts had different patterns compared to absorption spectra of low molecular weight

(MW<1KDa) fractions. High molecular weight coffee components, mostly latestage

MRPs produced absorbance throughout the whole wavelength spectra (250700 nm), but did not exhibit a clear peak absorbance pattern. This result indicates that numerous chromophores are presented in high the molecular weight coffee MRPs fractions. The finding is also supported by a previous study conducted with different sugaramino acids model MRPs (MacDougall and Granov, 1998). Low molecular weight coffee components

(MW<1KDa) that had two peaks at 275285 nm and 320 nm are similar to reports of low molecular weight MRPs that maximized at 280290 nm (Obretenov et al. , 1986; Bailey et

al. , 1996; Jing and Kitts, 2004a). Taken together, these results indicate that low molecular

67 weight coffee MRPs have distinct UV absorbance at around 280 nm, which may correlate

to a similar chemical group. Results from the present study also indicate that low molecular weight coffee MRPs are likely transformed or degraded further under severe

DR conditions.

2.4.2 Antioxidant activity of coffee extracts

Data from the present work demonstrated that both LR and DR coffee extracts contained components that had strong overall antioxidant properties, which can be mainly attributed to the low molecular weight components (MW<1KDa). Fractions III and IV derived from both LR and DR coffee extracts, containing compounds having molecular weight below

1KDa, and these contributed about 80 % of the total peroxyl radical scavenging activity,

and about 70 % of the total ABTS •+ scavenging ability. Reducing power of the defatted coffee extracts was also similar for LR (76 %) and DR (66 %). Our results agree with others reports using a linoleic acid peroxidation system, where 78 % of the overall inhibitory effect of coffee brew was found in the low molecular weight fraction

(MW<1KDa) (Somoza et al. , 2003). The LR coffee extract had a higher antioxidant activity compared to DR under the current roasting conditions, which can be attributed to the phenolics present in the coffee extracts that underwent a lower degree of degradation in the LR processing condition. The results from the three assays on the antioxidant activity of individual fractions were, however, not always consistent. For example,

Fractions III and IV showed the highest activity in ORAC assay, while Fractions II and

III had the highest TEAC values. This indicates that mixtures of chemically different

68 compounds present in coffee fractions could be behaving with characteristically different antioxidant mechanisms.

The formation of MRPs and the degradation of natural phenolics are two events during the roasting process of coffee which also directly relate to the final antioxidant activity of coffee. Studies by Morales’ group found that some low molecular weight compounds, particularly CGA in coffee, are ionically attached to high molecular weight melanoidins

(MW>10KDa), thus contributing mostly to the antioxidant activity of coffee melanoidins

(DelgadoAndrade and Morales, 2005; DelgadoAndrade et al. , 2005). In the present study, however, after release of the ionically attached compounds, no change or even an increased antioxidant activity of coffee melanoidins (MW>1KDa) was observed. Albeit

CGA can attach to the melanoidin structures through ionic interactions, a recent study found that CGA could also be covalently bound to melanoidins (Bekedam et al. , 2008d).

The later may be more predominant in the present study compared to ionic interactions.

In this case, little CGA would be released from the melanoidin structures following salt

treatment. Thus, no change in the antioxidant activity of coffee melanoidins was expected.

In addition, low molecular weight coffee constituents that have lower antioxidant activity

compared to melanoidins, or interfere the antioxidant action of melanoidins due to the

ionic bonding, would be released after salt treatment, therefore, resulting in higher

antioxidant activity for coffee melanoidins. A similar result has been reported for the

glucosephenylalanine model MRPs, after salt treatment, where melanoidins showed a

higher ABTS radical scavenging activity compared to melanoidins that had not

undergone salt treatment (RufianHenares and Morales, 2007c). In contrast, melanoidins

derived from other glucoseamino acid model systems showed deceased antioxidant

69 activity after salt treatment (RufianHenares and Morales, 2007c). These inconsistent

results from simple model systems indicate that the attachment of low molecular weight

MRPs to the melanoidin structures has a different influence on the antioxidant activity of

melanoidins, which could be either an enhanced or reduced activity.

2.4.3 Biological effects of coffee extracts

The primary aim of this study was to determine if coffee brews had an effect on gastrointestinal enterocyte viability using Caco2 cells. The MTT assay is an established method to quantify cellular growth and toxicity (Mosmann, 1983). Cells were treated for up to 72 h to achieve maximal potential toxic and biochemical effects. DR and LR coffee extracts exerted similar time and concentration dependent cytotoxic effects on Caco2 cells. Other workers have reported that treating Caco2 cells with drip coffee brews, prepared according to a standard procedure for 24 h, significantly reduced MTT response compared to fresh orange extract treated cells (Ekmekcioglu et al. , 1999). Associated with this observation was a 46 % reduction of cytochrome c reductase activity relatively to control (Ekmekcioglu et al. , 1999). A recent study using bovine aorta endothelial cells found that after a 24 h exposure of cells to a filter coffee and espresso coffee, the MTT cell viability reduced to 10 % and 19 % of the control, respectively (Hegele et al. , 2009).

Hydrogen peroxide (H 2O2) was detected in both filter and espresso coffee brews obtained from standardized procedures, and incubation of coffee under cell culture conditions, yielded an up to 11fold increase in H 2O2 content (Hegele et al. , 2009). These workers

suggested that H 2O2 is a major product in coffee inducing cell death in vitro (Hegele et

al. , 2009).

70 The effect of coffee on the antioxidant enzymes in Caco2 cells was investigated in the present study since coffee brews exhibited high antioxidant activity in many chemical assays. A concentration of 0.1 mg freezedried coffee/ml culture medium was used, at which no cytotoxic effect to Caco2 cells was observed for up to 72 h. Of interest, was the result that both CAT activity and GSH content in Caco2 cells were found to decrease after coffee treatment. CAT catalyzes the reduction of H 2O2 to water, thus normal

functions of CAT and GSH are important to maintain the intracellular oxidation

reduction status (redox). Redox is an important regulator of various genes that are

involved in pathologically conditions, where chronic prooxidant states either initiate or

exacerbate cell damage (Nath et al. , 1998; Waris and Ahsan, 2006). The inhibitory effect

of coffee brew on both CAT activity and reduced GSH level could be associated with the

MRPs since reduced activities of CAT and GR and GSH content has been reported in

lymphocytes after MRPs treatment (Yen et al. , 2002). In addition, Int407 cells when

exposed to modeled MRPs, also showed decreased activities of SOD, CAT, GR and GPX

(Jing and Kitts, 2004b). Decreased antioxidant enzyme activities and GSH content has

also been reported in the blood and liver of mice that were feed methylglyoxal, an

intermediate compound of the Maillard reaction (Choudhary et al. , 1997; Ankrah and

AppiahOpong, 1999). Moreover, the potential of coffee and MRPs to generate ROS has been proposed by several researchers (Yen et al. , 2002; Hegele et al. , 2009), which may

help explain the responses of Caco2 cells to coffee extracts employed in this study.

However, further studies are needed to elucidate how ROS are generated through

Maillard reaction and how MRPs in the food system are related to either beneficial or

harmful effects associated with generation or detoxification of ROS.

71

CHAPTER III

COFFEE CONSTITUENTS AND MODULATION OF OXIDATIVE STATUS IN

CACO-2 CELLS

EXPERIMENT IIa: IN VITRO AND CELLULAR IN VITRO ANTIOXIDANT

ACTIVITY OF MAILLARD REACTION PRODUCTS DERIVED FROM

COFFEE

EXPERIMENT IIb: IN VITRO AND CELLULAR IN VITRO ANTIOXIDANT

ACTIVITY OF MAILLARD REACTION PRODUCTS DERIVED FROM

SUGAR-SERINE MODEL SYSTEMS

72 3.1 Introduction

Great changes in the chemical composition and biological activity of green coffee beans

take place during the roasting process. These include the degradation of natural phenolic

compounds, generation of Maillard reaction products (MRPs), carbohydrate

caramelization, and pyrolysis of organic compounds (Guillot et al. , 1996; Belitz and

Grosch, 1999). The chemical and biological properties of coffee bean constitutes derived

from processing have not yet been completely elucidated.

Most of the physiological effects of coffee beverages have been attributed to caffeine, the

diterpenes kahweol and cafestol, and phenolic compounds (Lam et al. , 1987; Kitts and

Wijewickreme, 1994; Carrillo and Benitez, 2000; Ranheim and Halvorsen, 2005).

Relatively little is known about the health related effects associated with coffee MRPs

due to the difficulty in isolating and recovering these compounds from coffee brews.

Coffee has been shown to have strong antioxidant activity as evidenced by sequestering

metal prooxidant and scavenging free radicals, which is at least partially related to MRPs

(Borrelli et al. , 2002; Takenaka et al. , 2005). More recently, a low molecular weight

coffee MRP, Nmethylpridinium was shown to have strong chemopreventive effects

towards modulating Phase II enzymes both in vitro and in vivo (Somoza et al. , 2003).

This finding brought a new perspective to the bioactive properties of coffee MRPs and

the relationship with human health. Also, two recent studies demonstrated the presence of

hydrogen peroxide (H 2O2) in roasted coffee brews and Maillard reaction model systems, and the concentration of H 2O2 was shown to increase significantly when the brew was placed in cell culture conditions (Muscat et al. , 2007; Hegele et al. , 2009). This H 2O2 could lead to nuclear translocation of the transcription factor NFκB in macrophages,

73 which may be related to an immunomodulatory effects occurring in the gut. Hydrogen peroxide has also been suggested to be the major contributor to the cytotoxicity attributed to both coffee and MRPs. However, the presence of H2O2 in coffee and MRPs could have

an impact on the cellular oxidative status, which has not yet been investigated.

Although the contribution of natural phenolic compounds to the antioxidant capacity of

coffee cannot be ruled out, we hypothesized that MRPs contribute mostly to the

antioxidant activity of coffee; with low molecular weight coffee MRPs possessing higher

antioxidant activity than that of the high molecular weight MRPs. We also investigated

the in vitro effects of coffee MRPs on the intracellular redox environment of human

colon adenocarcinoma Caco2 cells in which the cells were exposed to extracts generated

from both green coffee, roasted coffee and model MRPs. Cellular antioxidant enzyme

activities and glutathione content were measured in this regard. In particular, the

expression of intestinal Caco2 genes that are involved in human oxidative stress and

antioxidant defense system were investigated. This research also evaluated possible protective effects of coffee MRPs against H2O2induced oxidative stress in Caco2 cells.

74 3.2 Materials and method

3.2.1 Preparation of coffee and Maillard reaction products (MRPs)

100% Coffea arabica was roasted in a commercial roaster at 204 ºC and 232 ºC for about

12 min to obtain light roasted (LR) and dark roasted (DR) coffee beans. Defatted coffee

extracts were made according to methods previously described in Chapter II. Green

coffee beans were treated with liquid nitrogen prior to grinding. The green bean (GB)

extract was prepared the same as the roasted coffee extracts.

Two model MRPs were prepared using arabinoseserine (AraSer) and sucroseserine

(SucSer) reactants. A solution of sugar (50 mM) and amino acid (50 mM) in distilled

water (100 ml) was freezedried and the mixtures obtained were placed into a beaker

(1000 ml), and then dryheated in the oven. Based on the roasting process used for the

coffee beans, two temperatures, namely 204 ºC and 232 ºC were used for LR and DR

MRPs, respectively. After 12 min heating, the beakers were removed from the oven and

allowed to cool to room temperature in desiccators. An aliquot (5 g) of the reaction

mixture was suspended in distilled water (250 ml) and stirred for 12 min. The solution

was then filtered (Whatman No.4) and the filtrate containing the watersoluble MPRs was

collected. The residue on the filter paper was washed with distilled water. All filtrates

were collected and freezedried to give the water soluble MRPs, which was used for the

fractionation experiments.

Ultrafiltration with ddH 2O was run on GB, LR and DR coffee defatted extracts, and

coffee model MRPs using the same procedure as described in Chapter II.

75 3.2.2 Chemical analyses and Antioxidant assays

The L, a, b and E Hunter scale color parameters and the browning of the roasted coffee

extracts and fractions were measured. The UV and fluorescence spectra of both roasted

coffee extracts and nonroasted coffee extracts were measured. In addition, the UV and fluorescence spectra of chlorogenic acid (CGA) and tannins (Brew UK Ltd. Salisbury) were measured.

The ORAC and TEAC results were calculated as Trolox equivalents (TE values). CGA was used as the reference for the RP assay. For making data comparisons, individual values were referenced to 1 gram of freezedried sample for each of the different coffee extracts and fractions. Antioxidant capacity and reducing power measures were compared between GB, LR and DR defatted extracts in order to evaluate the effect of the roasting processes.

The processes of chemical analyses and antioxidant assays were followed as described in

Chapter II section 2.2.4.

3.2.2.1 Quantification of dicarbonyl compounds

Dicarbonyl compounds were quantified by reversedphaseHPLC after derivatization

with 2,3 diaminonaphthalene (DAN; Sigma, St. Louis, MO, USA) (Chen and Kitts,

2008b). Defatted coffee extracts in 10 mM phosphate buffer (pH 7.4) were incubated

with DAN in the presence of 3,4hexanedione (Sigma, St. Louis, MO, USA) (internal

standard) overnight at 4 ºC. Aqueous solutions of standard glyoxal (SigmaAldrich,

Oakville, ON, Canada), methylglyoxal (Sigma, St. Louis, MO, USA), 3deoxyglucosone

(Toronto Research Chemicals Inc. Canada), and glucosone (Sigma, St. Louis, MO, USA)

76 were treated the same way as coffee samples. The reaction mixture was then extracted by

ethyl acetate and evaporated until dry under nitrogen gas. The extract was reconstituted in

methanol and injected into a Sphereclone ODS2 column (Phenomenex, Torrance, CA)

eluted with gradient acetonitrile (ACN) and 0.2% formic acid: 0–13 min, 28–45% ACN;

13–25 min, 45–85% ACN; 25–28 min, 85% ACN, with a flow rate of 0.8mL/min. The

quinoxaline derivatives were detected by diode array detector (265 nm) and fluorescent

detectors (excitation at 267 nm and emission at 503 nm).

3.2.3 Cellular in vitro Assay

Caco2 cells (passage 23 to 40) used in this chapter were cultured according to the

conditions and procedures described in Chapter II. For the MTT assay, cells were seeded

in 96well plates at a density of 1 × 10 6 cells/ml 24 h before tests. For enzyme activities, glutathione content as well as realtime RT PCR, cells were seeded in culture plates at a density of 2.5× 10 5 cells/ml and allowed to reach 100 % confluence before treatments.

Cells were treated with extracts and fractions derived from nonroasted and roasted coffee beans and model MRPs, respectively. Cells received no treatment were used as controls. Both treated and control cells were rinsed with icecold phosphate buffer saline

(PBS, pH 7.2) and scraped into a 1.5 ml tube for a freezethaw treatment. Some cells (no more than 1× 10 6) were kept in another 1.5 ml tube for realtime RTPCR. In some experiments, oxidative stress was induced by exposure of treated Caco2 cells to 5 mM

H2O2 solution in culture medium for 2 h. MTT response and cellular antioxidant enzyme activities and glutathione content were tested using the methods described in Chapter II.

77 3.2.4 Real-Time Quantitative Reverse Transcription PCR (RQ RT-PCR) Array

3.2.4.1 RNA isolation and cDNA preparation

RNA was isolated using RT 2 qPCRGrade RNA Isolated Kit (PA001, SABioscience,

Frederick, MD, USA). Caco2 cells were lysed with lysis and binding buffer and passed through a filter column to recover clear lysate before loading onto a RNA spin column for total RNA isolation. RNA was treated with RNasefree DNase and suspended in RNase free H 2O. RNA quality and quantity were determined using a NanoDrop spectrophotometer (NanoDrop Technology, Wilmington, USA). All RNA samples had

260/280 ratios ≥ 2.0 and 260/230 ratio ≥ 1.7. The same amount (1.0 g) of total RNA from every sample was used for first strand cDNA synthesis using RT 2 First Strand Kit

(C03, SABioscience, Frederick, MD, USA). RNA and cDNA were stored at 80 º C.

3.2.4.2 Realtime RTPCR

Target mRNA was quantified on BioRad iQ5 (BioRad, USA) using RT 2Profiler™ PCR

Array system (PAHS065A, Frederick, MD, USA). This RT 2Profiler™ PCR Array

contains genespecific primer sets for a thoroughly researched set of 84 genes relevant to

oxidative stress and antioxidant defense system in human. There are also five

housekeeping genes and three RNA and PCR quality controls (Table 3.1; also see Figure

1.4 for the layout of 96well PCR array). RTPCR was carried out in a 96well plate in a

total volume of 25 l/well, consisting RT 2 qPCR Master Mix, first strand cDNA and pre

dispensed genespecific primer. Realtime PCR was performed according to the user

manual: initial DNA denaturation at 95 ºC for 10 min, followed by 40 PCR cycles of

denaturation at 95 ºC for 15 sec and annealing/extension 60 ºC for 1 min. Reactions were

carried out in triplicate and data analysis using C t method.

78 Table 3.1 Human oxidative stress and antioxidant defense PCR array gene table Position GeneBank Symbol Description A01 NM_000477 ALB Albumin A02 NM_000697 ALOX12 Arachidonate 12lipoxygenase A03 NM_021146 ANGPTL7 Angiopoietinlike 7 A04 NM_001159 AOX1 Aldehyde oxidase 1 A05 NM_000041 APOE Apolipoprotein E A06 NM_004045 ATOX1 ATX1 antioxidant protein 1 homolog (yeast) A07 NM_004052 BNIP3 BCL2/adenovirus E1B 19kDa interacting protein 3 A08 NM_001752 CAT Catalase A09 NM_002985 CCL5 Chemokine (CC motif) ligand 5 A10 NM_005125 CCS Copper chaperone for superoxide dismutase A11 NM_007158 CSDE1 Cold shock domain containing E1, RNAbinding A12 NM_000101 CYBA Cytochrome b245, alpha polypeptide B01 NM_134268 CYGB Cytoglobin B02 NM_001013742 DGKK Diacylglycerol kinase, kappa B03 NM_014762 DHCR24 24dehydrocholesterol reductase B04 NM_175940 DUOX1 Dual oxidase 1 B05 NM_014080 DUOX2 Dual oxidase 2 B06 NM_004417 DUSP1 Dual specificity phosphatase 1 B07 NM_001979 EPHX2 Epoxide 2, cytoplasmic B08 NM_000502 EPX Eosinophil peroxidase B09 NM_021953 FOXM1 Forkhead box M1 B10 NM_197962 GLRX2 2 B11 NM_153002 GPR156 G proteincoupled receptor 156 B12 NM_000581 GPX1 Glutathione peroxidase 1 C01 NM_002083 GPX2 Glutathione peroxidase 2 (gastrointestinal) C02 NM_002084 GPX3 Glutathione peroxidase 3 (plasma) C03 NM_002085 GPX4 Glutathione peroxidase 4 (phospholipid hydroperoxidase) C04 NM_001509 GPX5 Glutathione peroxidase 5 (epididymal androgenrelated protein) C05 NM_182701 GPX6 Glutathione peroxidase 6 (olfactory) C06 NM_015696 GPX7 Glutathione peroxidase 7 C07 NM_000637 GSR Glutathione reductase C08 NM_000178 GSS Glutathione synthetase C09 NM_001513 GSTZ1 Glutathione transferase zeta 1 C10 NM_001518 GTF2I General transcription factor II, i C11 NM_006121 KRT1 Keratin 1 C12 NM_006151 LPO Lactoperoxidase D01 NM_000242 MBL2 Mannosebinding lectin (protein C) 2, soluble (opsonic defect) D02 NM_004528 MGST3 Microsomal glutathione Stransferase 3 D03 NM_000250 MPO Myeloperoxidase D04 NM_002437 MPV17 MpV17 mitochondrial inner membrane protein

79 Table 3.1 Human oxidative stress and antioxidant defense PCR array gene table (continued) Position GeneBank Symbol Description D05 NM_012331 MSRA Methionine sulfoxide reductase A D06 NM_005954 MT3 Metallothionein 3 D07 NM_004923 MTL5 Metallothioneinlike 5, testisspecific (tesmin) D08 NM_000265 NCF1 Neutrophil cytosolic factor 1 D09 NM_000433 NCF2 Neutrophil cytosolic factor 2 D10 NM_003551 NME5 Non metastatic cells 5, protein expressed in (nucleoside diphosphate kinase) D11 NM_000625 NOS2 Nitric oxide synthase 2, inducible D12 NM_024505 NOX5 NADPH oxidase, EF hand calcium binding domain 5 E01 NM_002452 NUDT1 Nudix (nucleoside diphosphate linked moiety X) type motif 1 E02 NM_181354 OXR1 Oxidation resistance 1 E03 NM_005109 OXSR1 Oxidativestress responsive 1 E04 NM_020992 PDLIM1 PDZ and LIM domain 1 E05 NM_015553 IPCEF1 Interaction protein for cytohesin exchange factors 1 E06 NM_007254 PNKP Polynucleotide kinase 3'phosphatase E07 NM_002574 PRDX1 Peroxiredoxin 1 E08 NM_005809 PRDX2 Peroxiredoxin 2 E09 NM_006793 PRDX3 Peroxiredoxin 3 E10 NM_006406 PRDX4 Peroxiredoxin 4 E11 NM_181652 PRDX5 Peroxiredoxin 5 E12 NM_004905 PRDX6 Peroxiredoxin 6 F01 NM_020820 PREX1 Phosphatidylinositol3,4,5trisphosphatedependent Rac exchange factor 1 F02 NM_006093 PRG3 Proteoglycan 3 F03 NM_183079 PRNP Prion protein F04 NM_000962 PTGS1 Prostaglandinendoperoxide synthase 1 (prostaglandin G/H synthase and cyclooxygenase) F05 NM_000963 PTGS2 Prostaglandinendoperoxide synthase 2 (prostaglandin G/H synthase and cyclooxygenase) F06 NM_012293 PXDN Peroxidasin homolog (Drosophila) F07 NM_144651 PXDNL Peroxidasin homolog (Drosophila)like F08 NM_014245 RNF7 Ring finger protein 7 F09 NM_182826 SCARA3 Scavenger receptor class A, member 3 F10 NM_203472 SELS Selenoprotein S F11 NM_005410 SEPP1 Selenoprotein P, plasma, 1 F12 NM_003019 SFTPD Surfactant protein D G01 NM_016276 SGK2 Serum/glucocorticoid regulated kinase 2 G02 NM_012237 SIRT2 Sirtuin (silent mating type information regulation 2 homolog) 2 (S. cerevisiae) G03 NM_000454 SOD1 Superoxide dismutase 1, soluble G04 NM_000636 SOD2 Superoxide dismutase 2, mitochondrial

80 Table 3.1 Human oxidative stress and antioxidant defense PCR array gene table (continued) Position GeneBank Symbol Description G05 NM_003102 SOD3 Superoxide dismutase 3, extracellular G06 NM_080725 SRXN1 Sulfiredoxin 1 homolog (S. cerevisiae) G07 NM_006374 STK25 Serine/threonine kinase 25 (STE20 homolog, yeast) G08 NM_000547 TPO Thyroid peroxidase G09 NM_003319 TTN Titin G10 NM_032243 TXNDC2 Thioredoxin domain containing 2 (spermatozoa) G11 NM_003330 TXNRD1 Thioredoxin reductase 1 G12 NM_006440 TXNRD2 Thioredoxin reductase 2 H01 NM_004048 B2M Beta2microglobulin H02 NM_000194 HPRT1 Hypoxanthine phosphoribosyltransferase 1 H03 NM_012423 RPL13A Ribosomal protein L13a H04 NM_002046 GAPDH Glyceraldehyde3phosphate dehydrogenase H05 NM_001101 ACTB Actin, beta H06 SA_00105 HGDC Human Genomic DNA Contamination H07 SA_00104 RTC Reverse Transcription Control H08 SA_00104 RTC Reverse Transcription Control H09 SA_00104 RTC Reverse Transcription Control H10 SA_00103 PPC Positive PCR Control H11 SA_00103 PPC Positive PCR Control H12 SA_00103 PPC Positive PCR Control

3.2.5 Statistical analysis

Each experiment was performed in triplicate (e.g. three wells or three cell culture plates)

and repeated three times in separate experiments. Collected data were expressed as mean

+ SD. Means were compared by Oneway Analysis of Variance (ANOVA), followed by

Tukey’s pairwise comparisons. Comparison of different treatments with control was done

by oneway ANOVA, followed by Dunnett’s test. Significant differences between two

samples were analyzed with Student ttest. Data with two influence factors were

compared by twoway ANOVA with Bonferroni posttests. The level of confidence

required for significance was selected at p<0.05. Statistical analyses were done by using

GraphPad Prism software (version 5.01, GraphPad Software, Inc.).

81 3.3 Results

Experiment IIa.

3.3.1 Recovery of coffee brews and fractions

Coffea arabica (100%) was roasted in a commercial roaster at 204 ºC and 232 ºC for

approximately 12 min to obtain light roasted (LR) and dark roasted (DR) coffee beans.

Lyophilized coffee brew samples were defatted and separated using water ultrafiltration.

The same procedure was performed on a coffee extract prepared from green coffee beans

(GB), taken as a control. For each coffee sample, four fractions (IIV) were collected and

the recovery for each fraction is reported in Table 3.2. GB coffee extract had greater

recovery of low molecular weight components (Fraction IV) and less high molecular

weight components (Fraction I), than those recovered from roasted coffee brew (p<0.05).

Fraction IV contained the highest recovery of components in LR coffee, while Fraction I

had the highest recovery in DR coffee. Fraction I recovered from LR and DR increased

upon roasting, while the recovery decreased in Fraction IV with more roasting (p<0.05).

Table 3.2 Recovery of coffee fractions by ultrafiltration 1 Fractions (MW) Recovery (% of defatted coffee dry weight) GB LR DR I (>10KDa) 15.0+0.6 ax 28.8+2.0 bcy 34.5+2.7 cy II (110KDa) 10.7+1.3 a 9.5+2.0 a 14.1+2.2 a III (0.51KDa) 25.7+3.6 b 23.5+4.3 b 20.6+7.0 ab IV (<0.5 KDa) 42.2+4.9 cy 32.6+6.8 cx 28.9+6.8 bcx 1 The values are expressed as mean + SD, n=3. abc means within the same column that do not share a common superscript letter are significantly different; xy represent significantly different means in rows. Statistical analyses using twoway ANOVA with Bonferroni posttests; statistical level of confidence set as 0.05. GB = green bean extract; LR = light roasted coffee extract; DR = dark roasted coffee extract.

82 3.3.2 Chemical characteristics of coffee

3.3.2.1 Colorimetric measurements of roasted coffee

Table 3.3 presents the colorimetric measurements reported on freezedried LR and DR

coffee extracts and related ultrafiltration fractions. Browning of green coffee beans

occurred after roasting due to the generation of polymerized MRPs. No significant difference in colorimetric parameter L was found between LR and DR defatted coffee extracts, however, it was observed that L values of individual fractions recovered increased as the molecular weight of components that were recovered decreased.

3.3.2.2 Browning of coffee extracts

An increase in browning corresponded to an increase in molecular weight of individual components (Table 3.3). This result implied a positive correlation between the degree of browning and the molecular weight of components present in derived products. No significant difference in browning was found between LR and DR.

Table 3.3 Lightness (L) and browning of coffee extracts and untrafiltration fractions 1 Coffee extracts L Browning 3 LR DR LR DR Defatted brew 34.0+4.8 40.2+1.6 0.19+0.02 0.18+0.00 Fractions (MW) 2 I (>10KDa) 26.6+1.7 a 23.2+0.6 a 0.33+0.03 d 0.33+0.02 c II (110KDa) 37.7+0.8 b 38.3+1.0 b 0.25+0.01 c 0.25+0.04 c III (0.51KDa) 38.5+1.7 bx 44.2+1.4 bcy 0.17+0.03 by 0.11+0.01 bx IV (<0.5 KDa) 43.2+2.8 b 45.6+1.4 c 0.08+0.00 a 0.07+0.00 a

1 LR and DR represent light roasted and dark roasted coffee. Values represent mean + SD, n = 3. abcd The means with different superscript letters in each column are significantly different; xy represent significantly different means in rows between LR and DR (p<0.05, twoway ANOVA with Bonferroni posttests). 2 Fractions are derived from defatted extracts, based on molecular weight (MW). 3 Browning intensities are absorbance reading at 420 nm. Samples were diluted to 0.5 mg/ml with distilled deionized water prior to spectrum measurement.

83 3.3.2.3 Fluoresence and UVvisible spectra

The GB, LR and DR coffee defatted extracts were also compared for relative difference

in fluorescence spectra (Figure 3.1). A sharp, distinct peak appearing at 400 nm emission

wavelength was observed in the GB coffee defatted extract and related fractions possessed (Figure 3.1B). Fraction IV recovered from GB defatted extract had a small peak occurring at 400 nm. The other three fractions recovered from the GB extract also had very high intensity peaks (Not shown in the Figure in order to avoid high overlapping). These peaks were summarized as follows: FI Fraction I : 1020; FI Fraction II : 860;

FI Fraction III : 1010 (where FI = fluorescent intensity). To verify that these peaks were related to natural phenolics, a fluorescence spectrum for chlorogenic acid (CGA) and tannins (0.05 mg/ml) was also measured. The concentration of CGA was based on an estimated proportion (e.g. 20 %) in soluble GB extracts (Farah and Donangelo, 2006;

Arya and Rao, 2007). The fluorescence spectra of 0.05 mg/ml CGA exhibited a sharp peak at 400 nm emission wavelength (Figure 3.1A). However, the intensity of this peak

(≈120) was much lower than the peak obtained from the GB coffee extracts. The peak of tannins (FI = 400) at 400 nm emission wavelength had an intensity that was relatively higher than CGA. This result indicated that high molecular weight polyphenolics are present in green coffee beans extract, and possessed components with a characteristic spectra pattern observed at a 400 nm emission wavelength. Fraction I recovered from both LR and DR coffee also contained a sharp fluorescent peak at 400 nm, and another broader peak at a higher emission wavelength (Figure 3.1D). The intensity of the peak at

400 nm emission wavelength from the roasted coffee was much lower than that derived from the GB coffee, likely indicating a loss of phenolic compounds during the roasting

84 process. The peaks at 400 nm present in Fractions II, III and IV recovered from roasted

coffee had relatively lower intensities compared to the broad peaks that occurred between

450 nm to 500 nm in these same fractions (Figure 3.1 EG). This finding supports the possibility that greater amounts of MRPs were presented in these three fractions than phenolics after roasting of the coffee beans. The pattern of the fluorescence spectra for

roasted coffee extracts shown here are similar to the fluorescent spectra for the roasted

coffee extracts in Chapter II (Figure 2.1).

Figure 3.2 shows the UV spectra of GB, LR and DR coffee defatted extracts and

ultrafiltration fractions over an absorbance range of 250700 nm. Both LR and DR

defatted coffee extracts had peak absorbance at around 275 nm, with a shoulder occurring

at 320 nm (Figure 3.2B). The UV absorbance spectra of Fraction I recovered from LR

and DR coffee extracts decreased continuously with increasing wavelength. In contrast,

the absorbance spectra for Fractions II and III recovered from LR and DR coffee extracts

had absorption peaks at 275 nm, with shoulders occurring at longer wavelengths. Fraction

IV also had a sharp peak at 275 nm. GB coffee samples had different UV spectral patterns compared to roasted coffee samples (Figure 3.2 BF). The UV absorbance

spectrum of 0.1 mg/ml CGA was also tested (Figure 3.2A). As expected, CGA showed a

similar UV spectral pattern as observed with the GB coffee extract. These results indicate

that GB contained relatively higher amounts of CGA, which greatly influenced the UV

spectral pattern of GB coffee extract. However, these phenolic compounds were altered

during roasting, resulting in products that did not possess a similar UV spectral pattern.

85

Figure 3.1 Fluorescence emission spectra (350550 nm) of green bean (GB), light roasted (LR) and dark roasted (DR) coffee extracts and fractions. The fluorescent spectra of 0.05 mg/ml CGA ( ) and tannins ( ) are shown in (A). The spectra of GB defatted extract ( ) and Fraction IV ( ) are shown in (B). LR ( ) and DR ( ) defatted coffee extracts (C), Fraction I (D), Fraction II (E), Fraction III (F) and Fraction IV(G). All emission spectra were measured with the excitation wavelength set at 400 nm. Samples were diluted to 0.25 mg/ml with miliQ water prior to spectrum measurement.

86

Figure 3.2 Comparison of the UVvisible spectra of green bean (GB), light roasted (LR) and dark roasted (DR) coffee extracts and fractions. The UV spectra of 0.1 mg/ml CGA ( ) and tannins ( ) are shown in (A). The spectra of GB ( ), LR ( ) and DR ( ) are defatted coffee extracts (B), Fraction I (C), Fraction II (D), Fraction III (E) and Fraction IV(F). Samples were diluted to 0.5 mg/ml with distilled deionized water prior to spectrum measurement.

87 3.3.2.4 Characterization of coffee αdicarbonyl compounds

Derivatization of αdicarbonyl compounds occurring in defatted coffee extracts was performed using 2,3diaminonaphthalene derivatization at 4 ºC overnight. Aqueous

solutions of standard glyoxal, methylglyoxal, 3deoxyglucosone, and glucosone were also

derivatized and 3, 4hexanedione was added as the internal control. Glyoxal and

methylglyoxal were identified in both defatted LR and DR coffee extracts. The

concentrations of these two specific αdicarbonyl are shown in Figure 3.3; methylglyoxal being the predominant αdicarbonyl compound recovered from both LR and DR coffee,

respectively. Methylglyoxal concentration in LR coffee was significantly higher (p<0.05)

than that in DR coffee (p<0.05), while there was no significant difference in the glyoxal

content between LR and DR coffee.

Figure 3.3 Alphadicarbonyl compounds in light roasted (LR: ) and dark roasted (DR: ) coffee extracts. * indicates significant difference between LR and DR (Student ttest, p<0.05).

3.3.3 Antioxidant activity of roasted and green coffee

The antioxidant capacity and reducing power of GB, LR and DR coffee extracts and

ultrafiltration fractions were measured by ORAC, TEAC and RP assays (Table 3.43.7).

88 Table 3.4 presents the affinity of GB, LR, and DR coffee to scavenge peroxyl radicals

and ABTS•+ radicals, respectively, and corresponding reducing activity. The antioxidant

indicator values for ORAC, TEAC and RP for GB defatted extract were significantly

higher (p<0.05) than those obtained from LR and DR coffee. Data from the ORAC and

TEAC assays suggested that the free radical scavenging activity decreased with increased

roasting processing of coffee beans.

Table 3.4 Antioxidant activity of coffee extracts 1 Coffee Assays 2 ORAC TEAC RP GB 2.50+0.09 c 0.63+0.01 c 0.31+0.02 b LR 1.39+0.07 b 0.53+0.02 b 0.16+0.00 a DR 1.12+0.06 a 0.45+0.01 a 0.15+0.01 a 1 Value represents mean + SD (n=3); abc represent significant different means in columns (p<0.05, oneway ANOVA with Tukey’s pairwise comparisons). GB = green bean extract; LR = light roasted coffee extract; DR = dark roasted coffee extract. 2 ORAC represents oxygen radical absorption capacity; TEAC represents Trolox equivalent antioxidant capacity; RP represent reducing power. ORAC and TEAC values are expressed as mmol Trolox equivalents/g freeze dried coffee samples; RP values are expressed as mg CGA/g freeze dried coffee samples.

Table 3.5 presents the ORAC values for GB, LR, and DR coffee fractions. Fraction III recovered from both GB and the LR coffee defatted extracts exhibited the highest antioxidant activity against generated peroxyl radicals (p<0.05). Fractions II and III recovered from DR coffee beans also showed higher ORAC activity compared to

Fractions I and IV (p<0.05). The ORAC values of GB fractions were significantly

(p<0.05) higher in comparison to the corresponding fractions recovered from roasted coffee. Distinct from ORAC results, Fractions II and III recovered from GB had the highest (p<0.05) TEAC values. Fraction II from both LR and DR coffee had the highest

89 (p<0.05) TEAC value (Table 3.6). Fractions II, III and IV recovered from GB had higher

TEAC values than those recovered from LR and DR coffee, respectively, while the

TEAC of LR Fraction I was higher than that of the GB Fraction I (p<0.05). Table 3.7 presents the ferric reducing power of coffee fractions. Among GB fractions, the RP of

Fractions II and III were significantly higher (p<0.05) than that of Fractions I and IV. No significant difference in RP was found among LR fractions. For DR coffee, Fraction II had the highest RP value, but this was not significantly different from Fractions I and III.

The RP values of GB coffee fractions were significantly higher (p<0.05) than those of

roasted coffee fractions.

Table 3.5 Antioxidant activity of coffee fractions determined by the ORAC method 1 Fractions (MW) 2 ORAC (mmol TE/ g sample) GB LR DR I (>10 KDa) 2.13+0.16 ay 1.31+0.01 ax 1.11+0.10 ax II (110 KDa) 2.51+0.11 az 1.67+0.01 by 1.39+0.06 bx III (0.51 KDa) 3.38+0.19 bz 1.98+0.02 cy 1.45+0.06 bx IV (<0.5 KDa) 2.37+0.27 az 1.38+0.03 ay 1.03+0.08 ax 1 ORAC represents oxygen radical absorption capacity. Values are expressed as mmol Trolox equivalents/g freeze dried coffee samples (mean + SD, n=3). abc means within the same column that do not share a common superscript letter are significantly different; xyz represent significantly different means in rows. Statistical analyses using twoway ANOVA with Bonferroni posttests; statistical level of confidence set at 0.05. GB = green bean extract; LR = light roasted coffee extract; DR = dark roasted coffee extract. 2 Fractions were derived from defatted coffee extracts, based on molecular weight (MW).

90 Table 3.6 Antioxidant activity of coffee fractions determined by the TEAC method 1 Fractions (MW) 2 TEAC (mmol TE/ g sample) GB LR DR I (>10 KDa) 0.44+0.03 ax 0.56+0.02 by 0.50+0.02 bcxy II (110 KDa) 0.84+0.01 cz 0.73+0.01 dy 0.56+0.02 cx III (0.51 KDa) 0.80+0.03 cz 0.66+0.03 cy 0.46+0.03 bx IV (<0.5 KDa) 0.68+0.07 bz 0.48+0.05 ay 0.39+0.02 ax 1 TEAC represents Trolox equivalent antioxidant capacity. Values are expressed as mmol Trolox equivalents/g freeze dried coffee samples (mean + SD, n=3). abc means within the same column that do not share a common superscript letter are significantly different; xyz represent significantly different means in rows. Statistical analyses using twoway ANOVA with Bonferroni posttests; statistical level of confidence set as 0.05. GB = green bean extract; LR = light roasted coffee extract; DR = dark roasted coffee extract. 2 Fractions were derived from defatted extracts, based on molecular weight (MW).

Table 3.7 Antioxidant activity of coffee fractions determined by the RP method 1 Fractions (MW) 2 RP (mmol CGA/ g sample) GB LR DR I (>10 KDa) 0.20+0.04 ay 0.17+0.01 x 0.17+0.01 abx II (110 KDa) 0.43+0.05 cy 0.22+0.02 x 0.21+0.02 bx III (0.51 KDa) 0.42+0.05 cy 0.19+0.01 x 0.15+0.01 abx IV (<0.5 KDa) 0.35+0.05 by 0.16+0.01 x 0.14+0.01 ax 1 RP represents reducing power. Values are expressed as mg CGA/g freeze dried coffee samples (mean + SD, n=3). abc means within the same column that do not share a common superscript letter are significantly different; xy represent significantly different means in rows. Statistical analyses using twoway ANOVA with Bonferroni posttests; statistical level of confidence set as 0.05. GB = green bean extract; LR = light roasted coffee extract; DR = dark roasted coffee extract. 2 Fractions were derived from defatted extracts, based on molecular weight (MW).

3.3.4 Biological effects of coffee bean extracts

3.3.4.1 MTT response

The MTT assay is based on the reduction of the tetrazolium ring of MTT by mitochondrial dehydrogenases yielding a purple formazan product, which is measured spectrophotometrically. The amount of formazan produced is proportional to the number

91 of viable cells. Caco2 cells were exposed to defatted extracts derived from GB, LR and

DR coffee samples with a concentration ranging from 0.005 to 10 mg freezedried extract/ ml culture medium. Cell response curves are presented in Figure 3.4 and the IC50 of Caco2 cells after treatment with GB, LR, and DR coffee defatted extracts for different time periods are summarized in Table 3.9. Both LR and DR coffee extracts reduced the viability of Caco2 cells in a time and concentrationdependent manner. GB extract did not show the same cytotoxicity as roasted coffee extracts. Significantly different (p<0.05)

IC50 were obtained for GB coffee treated cells, compared to cells treated with LR and DR

coffee. For all time periods, the IC50 values of roasted coffee treated cells were lower than cells treated with GB coffee (p<0.05). There was no significant difference in IC50 values between cells treated with LR and DR coffee extracts.

1 Table 3.9 IC 50 of coffee extracts on Caco2 cells using MTT assay

Time IC 50 (mg/ml) GB LR DR 3h 16.60+1.91 cx 13.57+1.30 cx 12h 21.81+1.17 ay 7.01+0.26 bx 6.98+0.35 bx 24h 22.28+1.34 ay 5.72+0.37 abx 5.28+0.13 ax 48h 31.73+1.25 cy 4.73+0.19 abx 4.54+0.14 ax 72h 26.95+1.19 by 3.99+0.16 ax 3.75+0.10 ax 1 IC 50 is determined as the concentration of coffee that is required to reduce cell viability to half that of control in the MTT assay. Values represent means + SD obtained for triplicate samples measured in triplicate. abcd represent means in columns that are significantly different; xy represent significantly different means in rows. Statistical analyses were done by oneway ANOVA with Tukey’s pairwise comparisons; level of confidence set as 0.05. GB = green beans; LR = light roast; DR = dark roast.

92 Figure 3.4 Effects of green bean (GB), dark roasted (DR) and light roasted (LR) coffee extracts on the tetrazolium reduction rate in the MTT assay after 3 h ( ), 12 h ( ), 24 h ( ), 48 h ( ) and 72 h ( ) incubation. (A): GB; (B): LR; (C): DR Data (% control) represent means + SD obtained for triplicate samples measured in triplicate.

93 3.3.4.2 Caco2 cellular glutathione content after coffee extracts treatment

All types of coffee used in this study showed high radical scavenging activity and

reducing capacity in the chemical antioxidant assays. However, the effects of coffee extracts in the biological system used herein produced different results. It has been reported that Caco2 cells exhibited antioxidant mechanisms mainly through the glutathione (GSH) cycle, a principle system that involves the adaptation to, and prevention of, cell oxidative damage (Baker and Baker, 1993). Considering the role of

GSH as one of the most important intracellular defenses, the GSH level in Caco2 cells exposed to 1.0 mg/ml of GB, LR and DR coffee defatted extracts for up to 72 h was measured (Figure 3.5). The concentration of coffee extracts selected from the MTT response was based on preliminary evidence that showed no cytotoxic effect at this concentration. The concentrations of 1.0 mg/ml and 2.0 mg/ml have been commonly used by other researchers to look at the cellular effects of coffee and other MRPs containing food (Somoza et al. , 2003; Jing and Kitts, 2004b; Muscat et al. , 2007). There was no significant difference in GSH content, quantified in cultures which were exposed to GB,

LR, and DR coffee extracts, respectively, for 3 h, compared to control. In contrast, cells treated with LR and DR coffee extracts, respectively, for 24 h and 72 h, exhibited a decreased intracellular GSH content relative to control (p<0.05), while no significant change in GSH content was found in cells treated with GB extract.

94

Figure 3.5 Effect of coffee extracts on glutathione (GSH) content in Caco2 cells. Cells were exposed to1.0 mg/ml green bean (GB), light roasted (LR) and dark roasted (DR) coffee for 3 h (A), 24 h (B) and 72 h (C). Cells incubated in culture medium were used as control. Results are mean + SD, n=3. * indicates significant difference compared to corresponding control (p<0.05). Statistical analysis was done by oneway ANOVA with Dunnett’s multiple comparison tests.

95 3.3.4.3 Caco2 antioxidant enzymes following exposure to coffee extracts

It is proposed that antioxidant enzymes are primarily responsible for the change in intracellular antioxidant status and are fairly sensitive biomarkers of cellular oxidative stress and activity of the antioxidant defense system. As shown in Figure 3.6, the activity of glutathione peroxidase (GPX) was significantly decreased in Caco2 cells (p<0.05) after 3 h exposure to 1.0 mg/ml LR and DR coffee defatted extracts. However, the activity of GPX increased significantly (p<0.05) after 24 h LR coffee exposure. Increased

GPX activity was also observed in DR coffee treated cells, but this change was not significant. Increased GPX activity in Caco2 cells could partially explain the reduced

GSH content observed at 24 h after LR coffee treatment, since increased GPX activity will accelerate the oxidation of GSH to GSSG. However, the reduction of GSSG in cells was limited due to the unchanged, or even slightly reduced, GR activity observed after 24 h of LR coffee treatment (not significant, p>0.05, data presented in Appendix Table 4).

After 72 h of LR coffee treatment, both GPX and CAT activity in Caco2 cells were significantly decreased (p<0.05), when compared to the control groups. The decreased

CAT activity was observed after 24 h treatment with LR and DR coffee extracts (p<0.05).

GB coffee treatments produced no stimulatory, or inhibitory, effects on the cellular antioxidant enzyme activities for up to 72 h exposure. This finding indicates a possible involvement of coffee MRPs in the regulation of antioxidant enzyme activity in Caco2 cells. Reduced superoxide dismutase (SOD) activities were observed in Caco2 cells exposed to coffee extracts, but the changes were not found to be significant (data in

Appendix Table 5).

96

Figure 3.6 Effect of coffee extracts on glutathione peroxidase (GPX) activity in Caco2 cells. Cells were exposed to1.0 mg/ml green bean (GB), light roasted (LR) and dark roasted (DR) coffee for 3 h (A), 24 h (B) and 72 h (C). Cells incubated in culture medium were used as control. Results are mean + SD, n=3. * indicates significant difference compared to corresponding control (p<0.05). Statistical analysis was done by oneway ANOVA with Dunnett’s multiple comparison tests.

97

Figure 3.7 Effect of coffee extracts on catalase (CAT) activity in Caco2 cells. Cells were exposed to1.0 mg/ml green bean (GB), light roasted (LR) and dark roasted (DR) coffee for 3 h (A), 24 h (B) and 72 h (C). Cells incubated in culture medium were used as control. Results are mean + SD, n=3. * indicates significant difference compared to corresponding control (p<0.05). Statistical analysis was done by oneway ANOVA with Dunnett’s multiple comparison tests.

98 3.3.5 Biological effects of coffee fractions on Caco-2 cells

Fractions II and III recovered from GB and LR coffee defatted extracts were chosen in this experiment in order to gain further insight into the biological effects of coffee containing components. Two different experimental designs were used: (1) treatment of

cells with 1.0 mg/ml coffee extracts and fractions, respectively, to test for a direct effect;

(2) pretreatment of cells with 1.0 mg/ml coffee extracts and fractions, respectively, before

exposing the cells to an oxidative stress using 5 mM H 2O2 for 2 h. Cell viability, GSH content, and antioxidant enzyme activities were evaluated in both experiments.

3.3.5.1 MTT response

Cellular MTT responses to GB and LR coffee samples, including defatted extracts,

Fractions II and III, both with (+) or without () H2O2 treatment are summarized in Table

3.10. After 24 h treatment of Caco2 cells with Fraction II recovered from LR coffee, a significant decrease (p<0.05) in cell viability was observed compared to that of the negative control, which contained Caco2 cells grown in culture medium without H2O2

treatment. No significant cytotoxic effect on Caco2 cells was found for other GB and LR

coffee samples after 3 h incubation. Increasing the exposure time of cells to coffee

samples to 24 h, did not produce any further significant stimulatory or inhibitory effect

on cellular MTT response.

Exposure of Caco2 cells to H 2O2, reduced cell viability to 80 % of negative control, and

this effect was not prevented by preincubation of cells with either GB or LR coffee

samples for 3 h or 24 h, respectively. The pretreatment of cells with defatted extract and

Fraction II derived from LR coffee resulted in a greater degree of Caco2 cytotoxicity,

compared to the positive control (+H2O2). Cell viability decreased to 63 % and 58 % of

99 negative control, respectively. In general, these results indicated that exposure of Caco2 cells to GB and LR coffee was not cytotoxic. The exception to this was when cells were

treated with Fraction II derived from LR coffee for 24 h. Moreover, these coffee

treatments of Caco2 cells did not produce a protective effect against H 2O2 induced cytotoxicity. In fact, exposure of cells to the defatted extract and Fraction II derived from

LR coffee may have further enhanced the cytotoxic effects of H 2O2.

1 Table 3.10 Caco2 MTT response to coffee with (+) and without () H 2O2 treatment

H 2O2 + H 2O2 Control 2 100+6 80+4* 3 h coffee Treatment GB H 2O2 + H 2O2 Defatted 101+3 90+6 Fraction II 110+1 93+3 Fraction III 93+5 92+2 LR Defatted 98+4 79+7 Fraction II 98+7 73+1 Fraction III 94+2 86+7 24 h coffee Treatment GB H 2O2 + H 2O2 Defatted 104+2 69+3 Fraction II 107+6 72+5 Fraction III 110+4 72+7 LR Defatted 92+1 63+2** Fraction II 80+4* 58+2** Fraction III 95+8 77+4 1 Data (% control) represent means + SD obtained for triplicate samples measured in triplicate. Significant changes are expressed as * (p<0.05) in comparison with negative ** control ( H2O2), and (p<0.05) in comparison with positive control (+ H2O2) (oneway ANOVA followed by Dunnett’s multiple comparison tests). 2 The value of positive control (+ H2O2) was significantly (p<0.05) lower than that of negative control ( H2O2).

100 3.3.5.2 Caco2 Cellular glutathione content following exposure to coffee extracts

The concentration of GSH was measured in Caco2 cells treated with defatted extracts,

Fractions II and III, respectively, derived from GB and LR coffee for 24 h. GSH content

was also measured in these cells which were subsequently treated with H 2O2 for 2 h. As a positive control to stimulate a prooxidant effect on changes in GSH utilization, cells were exposed to only H2O2 for 2 h and the results are shown in the same figure (Figure 3.8).

Significant decreases (p<0.05) in GSH content were observed in cells treated with defatted LR coffee extract and the recovered Fractions II, respectively, compared to the negative control. No significant change in GSH content was found in Caco2 cells after

GB sample treatment. Exposure of Caco2 cells to GB and LR coffee samples did not prevent H2O2induced reduction of cellular GSH content.

Figure 3.8 Effect of light roasted (A) and green bean (B) coffee extracts on Caco2 cellular GSH contents with (+: ) and without (: ) H 2O2 treatment. Caco2 cells treated with coffee samples alone were compared to negative control ( H2O2), and cells pretreated with coffee samples and further challenged by H2O2 were compared to positive control (+ H2O2). *represents significant difference in comparison with negative contol. There was no significant change in GSH content in cells that were pretreated with coffee samples and further challenged by H2O2 compared to that of positive control. Statistical analyses were done by oneway ANOVA with Dunnett’s tests (p<0.05).

101 3.3.5.3 Caco2 antioxidant enzyme activities following exposure to coffee extracts

The presence of H 2O2 in the culture medium induced a significant increase (p<0.05) in the enzyme activity of GPX compared to negative control (Figure 3.9 AB). Treating

Caco2 cells with the LR coffee extract alone, also increased GPX activity (p<0.05)

(Figure 3.9 A). The GPX induction was not observed when cells were treated with either

Fraction II or III derived from the LR coffee. No significant change in GPX activity was observed for cells treated with GB coffee samples alone (Figure 3.9 B). The GPX activity in cells pretreated with LR coffee samples was not significantly different compared to the positive control. However, cells pretreated with GB samples showed significantly

(p<0.05) decreased GPX activity compared to the positive control, indicating the prevention of H2O2 induced GPX activity when preexposed to GB extracts.

Exposure of Caco2 cells to H 2O2 resulted in a reduced cellular GR activity in the present

study (Figure 3.9 CD). Cells pretreated with LR coffee Fraction III exhibited a significant

decrease in the GR activity, compared to the positive control (p<0.05). No significant

changes in GR activity were found in all other samples when treated with coffee extracts

alone, or when producing a further induction in oxidative stress.

As shown in Figure 3.9 E and F, respectively, the presence of H 2O2 for 2 h resulted in a

significant increase in CAT activity in Caco2 cells (p<0.05). A pretreatment of cells with

LR coffee samples prevented the H2O2induced increase in CAT activity. Treatment of

Caco2 cells with either Fraction II or III derived from LR coffee had no effect on cellular CAT activity. There was no significant difference in Caco2 CAT activity between GB treated cells and corresponding control cells (p>0.05).

102 Caco2 SOD activity was not significantly affected in any of the different experimental conditions or treatments used above (data in Appendix Table 67).

103

Figure 3.9 Effect of light roasted (LR) and green bean (GB) coffee extracts on Caco2 cellular antioxidant enzyme activities with ( ) and without ( ) H 2O2 treatment. A: GPX activity, LR treatment; B: GPX activity, GB treatment; C: GR activity, LR treatment; D: GR activity, GB treatment; E: CAT activity, LR treatment; F: CAT activity, GB treatment. Cells treated with coffee samples alone were compared to negative control (H2O2), and cells pretreated with coffee samples and further challenged by H2O2 were compared to positive control (+H2O2). *indicates significant difference in comparison with negative control, and ** indicates significant difference in comparison with positive control. Statistical analyses were done by oneway ANOVA with Dunnett’s tests (p<0.05).

104 3.3.6 The regulatory effects of coffee on the expression of the genes involved in the

oxidative stress and antioxidant defense system in Caco-2 cells

Treating Caco2 cells with LR and DR defatted coffee extracts for 24 h resulted in both

an over and underexpression of numerous genes that are involved in the human

oxidative stress and antioxidant defense system. Table 3.11 summarizes the gene

expression responses obtained from Caco2 cells treated with GB, LR, and DR coffee extracts and hydrogen peroxide for 24 h; the latter of which was used as a positive control.

Of the three different coffee extracts tested, the one derived from green coffee beans, was used as the negative control. Changes in gene expression are expressed as a fold change, which is the ratio of the gene expression between treated and untreated conditions. A significant change in gene expression was denoted as a fold change greater than 2fold

(up or down) (p<0.05).

Exposure of Caco2 cells to GB coffee extract resulted in a significant (p<0.05) increase in GPX2 and a decrease (p<0.05) in MT3 expression. This result compared to 9 genes that

were significantly affected when cells were exposed to H 2O2 (p<0.05). Caco2 cells exposed to the DR coffee extract had 13 specific genes where expression was significantly altered (P<0.05). LR coffee extract exhibited the greatest influence on Caco

2 cell gene expression, with 11 genes being upregulated (p<0.05). Relative changes in gene expression after exposure to LR coffee extract were found to be greatest for GPX2

(+18.74fold) and iNOS (+11.46fold). Examples of genes that were downregulated, included NME5 (9.46fold), ALB (6.72fold), PRG3 (6.00fold), SEPP1 (5.73fold), and MT3 (5.21fold). Descriptions of these genes and relative function are given in the

Appendix Table 8. Genes containing antioxidant response elements (ARE) in the

105 promoter region are of particular interest in the present study. The regulation of these

genes by LR coffee extract, DR coffee extract, and H 2O2, respectively, are shown in the

Figure 3.10.

Overall, treatment of Caco2 cells with the LR coffee extract resulted in greater changes

in gene expression, both in the number of genes and more so in the magnitude of change

when compared to cells exposed to either DR coffee extract or H2O2, respectively. For

example, GPX2 was upregulated 18.74fold by LR coffee treatment and 13.88fold by

DR coffee treatment. In contrast, GPX2 expression was upregulated 2.43fold and 3.68

fold, when cells were treated with GB coffee (negative control) and H2O2 (positive control), respectively. iNOS expression increased significantly (p<0.05) in Caco2 cells treated with LR and DR coffee extracts and H2O2, respectively, but not in cells treated with GB extract. The gene regulatory effects of LR, DR and GB coffee extracts and H2O2

on the expression of GPX2 , iNOS , SRXN1 , TXRD1 , PRDX4 and CAT in Caco2 cells fell

into the same order: LR > DR > H 2O2 > GB; where cells treated with LR coffee extracts

had the greatest change in these three gene expression, followed by cells treated with DR

coffee extract and H2O2, respectively. No significant change occurred in the expression of these six genes when cells were treated with GB coffee extract.

106 Table 3.11 Genes differently expressed in Caco2 cells after incubation with coffee 1 extracts and H 2O2 Gene symbol Treatment GB LR DR H2O2 GPX2 +2.43 +18.74 +13.88 +3.68 iNOS n/c +11.46 +7.10 +2.58 SRXN1 n/c +5.00 +2.79 n/c NCF2 n/c +2.88 n/c n/c TXNRD1 n/c +2.84 +2.07 n/c PXDN n/c +2.79 n/c n/c PRDX4 n/c +2.46 +2.00 n/c MBL2 n/c +2.45 +2.64 n/c SELS n/c +2.29 n/c n/c PRDX1 n/c +2.10 n/c n/c PRDX6 n/c n/c n/c +2.00 NME5 n/c 9.46 n/c 2.42 ALB n/c 6.72 2.97 5.66 PRG3 n/c 6.00 2.75 n/c SEPP1 n/c 5.73 4.22 2.20 NOX5 n/c 5.25 4.95 7.77 MT3 2.85 5.21 2.15 4.11 CAT n/c 3.82 2.49 n/c PREX1 n/c 3.03 n/c n/c ATOX1 n/c 2.59 n/c n/c MSRA n/c 2.35 2.52 n/c AOX1 n/c 2.28 n/c 5.65 PTGS1 n/c 2.26 n/c n/c 1 Data represent fold changes of differently expressed genes in treated cells compared to control cells (n=3, fold change>2, p<0.05). “+” = upregulation; “” = downregulation; n/c = no significant change (p>0.05) or fold change<2. The name and function of these genes are described in Appendix (Table 8).

107

Figure 3.10 Antioxidant genes expression in Caco2 cells treated with light roasted (LR:

), dark roasted (DR: ) coffee extracts and H 2O2 ( ) compared to those in control cells. A: upregulated gene; B: downregulated gene. *** indicates p<0.001, ** indicates p<0.01, * indicates p<0.05. GPX2 = glutathione peroxidase 2; SRXN1 = sulfiredoxin 1; TXNRD1 = thioredoxin reductase 1; PRDX4 = peroxiredoxin 4; PRDX1 = peroxiredoxin 1; PRDX6 = peroxiredoxin 6; CAT = catalase.

108 EXPERIMENT IIb

3.3.7 Chemical characteristics and antioxidant activity of Suc-Ser and Ara-Ser model

MRPs

Studies have indicated that serine and threonine represent two amino acids that react with sugars, such as arabinose and sucrose to produce MRPs during the roasting process of green coffee beans (De Maria et al., 1996; Reichardt et al., 2009). In this experiment, arabinoseserine (AraSer) and sucroseserine (SucSer) MRPs were prepared according to the same roasting condition used to obtain LR and DR coffee beans, characterized earlier in Experiment II. The chemical characteristics and antioxidant activity associated with these two types of model MRPs were investigated in order to choose the appropriate model MRPs to represent MRPs derived from coffee beans.

3.3.7.1 Colorimetric measurements and browning of SucSer and AraSer model MRPs

Table 3.12 presents data of lightness (L) and browning obtained from SucSer and Ara

Ser model MRPs. L values for LR and DR AraSer MRPs were lower (p<0.05) than the corresponding L values for SucSer MRPs. These findings indicated greater darkness generated from heat processed AraSer reactions. Similarly, AraSer MRPs had a significantly higher (p<0.001) degree of browning, compared to SucSer MRPs when processed under identical conditions. There was no significant different in L values and browning between MRPs derived from LR and DR processes.

109 Table 3.12 Lightness (L) and browning of model MRPs 1 MRPs samples L Browning 2 LR DR LR DR SucSer MRPs 26.2+2.1 b 25.5+1.8 b 0.25+0.04 a 0.31+0.02 a AraSer MRPs 17.6+1.7 a 20.2+0.7 a 0.56+0.01 b 0.54+0.03 b

1 LR and DR represent light roasted and dark roasted MRPs. Values represent mean + SD, n = 3. ab The means with different superscript letters in each column are significantly different (p<0.05, twoway ANOVA with Bonferroni posttests). 2 Browning intensities are absorbance reading at 420 nm. Samples were diluted to 0.5 mg/ml with distilled deionized water prior to spectrum measurement.

3.3.7.2 Fluorescence and UV spectra of SucSer and AraSer model MRPs

Figure 3.11 shows the different fluorescence intensities obtained for SucSer and AraSer model MRPs. The emission spectra, with maximum emission wavelength at 470 nm for

SucSer MRPs and AraSer MRPs crude extracts were both obtained at an excitation wavelength at 400 nm. The maximum fluorescence intensity of DR MRPs derived from the SucSer model system was relatively greater (p<0.05) than that of the LR MRPs.

There was no significant difference in maximum fluorescence intensity between LR and

DR MRPs derived from the AraSer model system. However, LR MRPs derived from the

AraSer system had a peak that occurred at 400 nm emission wavelength, where the DR

AraSer MRPs had only little fluorescence.

Figure 3.12 shows the different UV spectra obtained for SucSer and AraSer model

MRPs, respectively. SucSer and AraSer MRPs crude extracts had similar characteristic peak absorbance at 285 nm, with the DR model systems showing relatively higher peak absorbance than the LR ones.

110

Figure 3.11 Fluorescence emission spectra (350550 nm) of light roasted (LR ) and dark roasted (DR ) SugarSerine MRPs extracts. SucSer MRPs extracts (A), and AraSer MRPs extracts (B), measured with the excitation wavelength set at 400 nm. Samples were diluted to 0.25 mg/ml with MilliQ water prior to spectrum measurement.

Figure 3.12 UV spectra of light roasted (LR ) and dark roasted (DR ) Sugar Serine MRPs extracts. SucSer MRPs extracts (A), and AraSer MRPs extracts (B), measured over a range of absorbance wavelength (250700 nm). Samples were diluted to 0.5 mg/ml with distilled deionized water prior to spectrum measurement.

111 3.3.7.3 Generation of αdicarbonyl compounds in model MRPs

Two αdicarbonyl compounds, glyoxal and methylglyoxal, were identified in both the LR

and DR AraSer MRPs (Figure 3.13 A). There was no significant difference in the

relative concentrations of αdicarbonyls between LR and DR MRPs derived from the

AraSer model system. Unlike AraSer MRPs, four αdicarbonyl compounds, namely

glyoxal, methylglyoxal, 3deoxyglucosone, and glucosone, were identified in the SucSer

model MRPs (Figure 3.13 B). The concentration of 3deoxyglucosone was significantly

higher (p<0.05) than that of the other three αdicarbonyl compounds, in both LR and DR

SucSer MRPs. The concentrations of glyoxal and methylglyoxal were higher in DR Suc

Ser MRPs compared to those of LR SucSer MRPs, respectively (p<0.05). In contrast, the

glucosone concentration was higher in the LR SucSer MRPs (p<0.05). In general, the

concentration of total αdicarbonyl compounds in SucSer MRPs was greater (p<0.05)

than that in AraSer MRPs. This was primarily due to the high 3deoxyglucosone content

in SucSer MRPs.

114

Figure 3.13 Alphadicarbonyl compounds in AraSer MRPs (A) and SucSer MRPs (B). Light roasted (LR: ); dark roasted (DR: ) MRPs crude extracts. DOG = 3deoxyglucosone; GLO = glyoxal; MGL = methylglyoxal; GLS = glucosone. * indicates significant difference between LR and DR (Student ttest, p<0.05).

115 3.3.7.4 Antioxidant activity of model MRPs

Table 3.13 summarizes the antioxidant activity and the reducing power measured from both the AraSer and SucSer MRPs crude extracts, using the ORAC, TEAC and the RP

assay. Generally, AraSer MRPs had relatively higher antioxidant activity compared to

SucSer MRPs (p<0.05).

Table 3.13 Antioxidant activity of SucSer and AraSer MRPs 1 MRPs Assays 2 ORAC TEAC RP LR DR LR DR LR DR SucSer 0.49+0.01 a 0.45+0.06 a 0.16+0.01 a 0.19+0.01 a 0.15+0.01 a 0.17+0.01 a AraSer 1.06+0.17 b 0.96+0.12 b 0.65+0.03 b 0.57+0.04 b 0.21+0.02 b 0.24+0.03 b 1 LR and DR represent light roasted and dark roasted model MRPs. Value represents mean + SD (n=3). ab represent significantly (p<0.05) different means in each column; no significant difference between LR and DR (p>0.05). Significant differences were analyzed using twoway ANOVA with Bonferroni posttests. 2 ORAC represents oxygen radical absorption capacity; TEAC represents Trolox equivalent antioxidant capacity; RP represents reducing power. ORAC and TEAC values are expressed as mmol Trolox equivalents/g freezedried MRPs samples; RP values are expressed as mg CGA/g freezedried MRPs samples.

MRPs derived from AraSer model system were chosen to represent coffee MRPs. The

reasons behind this decision were: (1) the αdicarbonyl compounds identified in AraSer

model MRPs were the same as those identified in coffee extracts. (2) The antioxidant

activity and reducing power of MRPs derived from AraSer system were significantly

higher compared to MRPs recovered from SucSer model system. (3) The antioxidant

activity of MRPs derived from AraSer MRPs extract was similar to that of coffee extract.

Crude extracts derived from AraSer model system were further fractionated based on

molecular weight and four fractions were obtained.

116 3.3.8 Chemical characteristics and antioxidant activity Ara-Ser model MRPs fractions

3.3.8.1Recovery of fractions derived from AraSer model system

Multiplestep ultrafiltration was conducted to separate MRPs recovered from AraSer

model system into different molecular weight (MW) fractions, as reported earlier (e.g.

Fraction I (MW>10KDa), Fraction II (1KDa

(0.5KDa

contained the highest recovery of MRPs in both LR and DR processed AraSer reactants

(Table 3.14).

Table 3.14 Recovery of AraSer MRPs ultrafiltraton fractions 1 Fractions (MW) 2 Recovery (% of crude MRPs extract) LR DR I (>10KDa) 24.0+3.5 b 22.7+3.3 b II (110KDa) 7.7+1.4 ax 16.5+4.1 aby III (0.51KDa) 21.3+3.2 by 12.5+2.6 ax IV (<0.5 KDa) 41.8+3.8 c 44.3+6.3 c 1 Values are expressed as mean + SD, n=3. abc represent significantly different means in each column; xy represent significantly different means in rows. Statistical analyses using twoway ANOVA with Bonferroni posttests; statistical level of confidence set as 0.05. 2 Fractions were derived from AraSer MRPs extracts, based on molecular weight (MW).

3.3.8.2 Colorimetric measurements and browning of AraSer model MRPs

Table 3.15 presents data on the lightness (L) and browning of ultrafiltration fractions recovered from AraSer MR system. Fractions I, II and III recovered from AraSer model

MRPs showed greater darkness than Fractions IV (p<0.05).There was a pattern of relative increase in browning that corresponded to an increase in molecular weight in the four fractions recovered from AraSer MRPs.

117 Table 3.15 Lightness (L) and browning of fractions derived from AraSer model MR system 1 Fractions (MW) 2 L Browning 3 LR DR LR DR I (>10KDa) 20.9+0.1 a 21.0+3.1 a 0.98+0.06 c 1.20+0.29 d II (110KDa) 22.9+1.8 a 25.0+1.4 ab 0.92+0.07 c 0.96+0.07 c III (0.51KDa) 27.4+4.8 a 31.9+2.9 b 0.50+0.09 b 0.47+0.05 b IV (<0.5 KDa) 47.0+4.5 b 42.2+8.8 c 0.12+0.02 a 0.18+0.06 a

1 LR and DR represent light roasted and dark roasted MRPs. Values represent mean + SD, n = 3. abcd The means with different superscript letters in each column are significantly different. (p<0.05, twoway ANOVA with Bonferroni posttests). 2 Fractions are derived from AraSer MRPs extracts, based on molecular weight (MW). 3 Browning intensities are absorbance reading at 420 nm. Samples were diluted to 0.5 mg/ml with distilled deionized water prior to spectrum measurement.

3.3.8.3 Fluoresence and UV spectra

Figure 3.14 shows the different fluorescence intensities obtained for crude extract and ultrafiltration fractions derived from AraSer MRPs, respectively. Maximum fluorescence intensity was found at 480 nm for Fractions I and II; 470 nm for Fraction III, and 460 nm for Fraction IV. Among the four ultrafiltration fractions derived from AraSer MRPs extract, Fraction III gave the highest fluorescence intensity (p<0.05). The peak at 400 nm emission wavelength that was presented in the spectra for the LR AraSer MRPs crude extract was not observed in any of the four fractions recovered from the AraSer crude extract.

Figure 3.15 shows the different UV spectra obtained for AraSer model MRPs crude extract, as well as recovered ultrafiltration fractions. The UV spectral absorbance of Ara

Ser Fractions I, II and III were found to continuously decrease as the wavelength increased. Fraction IV recovered from AraSer MRPs had a peak UV absorbance at 285 nm.

118

Figure 3.14 Fluorescence emission spectra (350550 nm) of light roasted (LR ) and dark roasted (DR ) MRPs extracts and fractions. AraSer MRPs crude extracts (A), Fraction I (B), Fraction II (C), Fraction III (D) and Fraction IV(E), measured with the excitation wavelength set at 400 nm. Fractions were derived from AraSer MRPs extracts, based on molecular weight (MW). Samples were diluted to 0.25 mg/ml with MilliQ water prior to spectrum measurement.

119

Figure 3.15 Comparison of the UV spectra of light roasted (LR ) and dark roasted (DR ) MRPs extracts and fractions. AraSer MRPs crude extracts (A), Fraction I (B), Fraction II (C), Fraction III (D) and Fraction IV(E), measured over a range of absorbance wavelengthes (250700 nm). Fractions were derived from AraSer MRPs extracts, based on molecular weight (MW). Samples were diluted to 0.5 mg/ml with distilled deionized water prior to spectrum measurement.

120 3.3.8.4 Antioxidant activity of AraSer model MRPs fractions

Table 3.16 summarizes the antioxidant activity measured on different ultrafiltration fractions collected from the model AraSer MRPs extracts. Fractions I and III recovered from both the LR and DR AraSer MRPs, possessed higher ORAC antioxidant activity compared to Fractions II and IV (p<0.05). Higher radical scavenging activity was observed in Fractions I and II for both LR and DR AraSer MRPs when using the TEAC assay. No significant difference in RP activity was found between all fractions tested

(data in Appendix Table 9).

Table 3.16 Antioxidant activity of fractionated AraSer MRPs 1 Fractions 2 Assays 3 ORAC TEAC LR DR LR DR I (>10 KDa) 1.48+0.05 c 1.46+0.09 c 0.88+0.07 c 0.88+0.03 c II (110 KDa) 0.87+0.03 b 0.88+0.01 b 0.80+0.06 c 0.83+0.06 c III (0.51KDa) 1.54+0.04 c 1.51+0.15 c 0.49+0.05 b 0.47+0.04 b IV (<0.5 KDa) 0.39+0.03 a 0.52+0.05 a 0.20+0.02 a 0.24+0.06 a 1 LR and DR represent light roasted and dark roasted model MRPs. Value represents mean + SD (n=3). abc The means with different superscript letters in each column are significantly different (p<0.05); no significant difference between LR and DR. Significant differences were analyzed using twoway ANOVA with Bonferroni posttests. 2 Fractions were derived from AraSer MRPs extracts, based on molecular weight (MW). 3 ORAC represents oxygen radical absorption capacity; TEAC represents Trolox equivalent antioxidant capacity. ORAC and TEAC values are expressed as mmol Trolox equivalents/g freezedried MRPs samples.

3.3.10 Cell-based bioactivity of Ara-Ser MRPs

3.3.10.1 Caco2 MTT response

MRPs recovered from heat processed AraSer reactants to generate model MRPs similar to that recovered from roasted coffee were found to reduce the viability of Caco2 cells in both a concentration and timedependent manner. The concentration response curves of

121 Caco2 cells after exposure to both LR and DR model MRP treatments are shown in

Figure 3.16. Calculated IC50 values for cells treated with these model MRPs at different time periods are presented in Table 3.17. The result showed that a concentration of 1.0 mg/ml, similar to the concentrations of coffee samples used in experiment IIa, should be chosen for further cellbased antioxidant experiments in order to compare the results to the coffee MRP data. It should be point out that LR and DR MRP at a concentration of

1.0 mg/ml treatment of Caco2 cells for 72 h reduced cell viability by about 25%. This observed cytotoxicity was significant greater (p<0.05) compared to the MRPfree control.

122

Figure 3.15 Effects of light roasted (LR) and dark roasted (DR) AraSer MRPs extracts on the tetrazolium reduction rate in the MTT assay after 3 h ( ), 12 h ( ), 24 h ( ), 48 h ( ) and 72 h ( ) incubation. A: LR; B: DR. Data (% control) represent means + SD obtained for triplicate samples measured in triplicate.

123 1 Table 3.17 IC 50 values of AraSer MRP extracts on Caco2 cells using MTT assay

IC 50 (mg/ml) LR DR 3h 13.72+0.59 d 14.20+0.84 e 12h 13.85+0.99 dy 9.41+0.23 dx 24h 10.26+0.24 cy 6.44+0.36 cx 48h 6.40+0.22 by 4.90+0.14 bx 72h 3.41+0.34 ay 1.94+0.14 ax 1 IC 50 is equal to the concentration of coffee that was required to reduce cell viability to half that of control in the MTT assay. Values represent means + SD obtained for triplicate samples measured in triplicate. abcde represent means in columns that are significantly different; xy represent significantly different means in rows. Statistical analyses were done by twoway ANOVA with Bonferroni posttests; level of confidence set at 0.05.

3.3.10.2 Caco2 cellular glutathione content

A significant (p<0.05) reduction of GSH content was observed in Caco2 cells after a 24h

treatment with LR and DR AraSer MRPs, respectively (Figure 3.16). No significant

change in GSH content was seen after relatively early (3 h) treatment, or after prolonged

(72 h) treatments. The in vitro cytotoxic effect of MRPs on Caco2 cells was not

associated with an intracellular reduction of GSH content in the present study, since 24 h

exposure of cells to 1.0 mg/ml MRPs did not correspond to a significant increase in

cytotoxicity.

3.3.10.3 Caco2 antioxidant enzyme activity

Exposing Caco2 cells to 1.0 mg/ml LR and DR AraSer model MRPs, respectively, did

not result in significant changes in the activity for CAT and GR up to 72 h (data in

Appendix Table 1011). Reduced cellular GPX activity was observed after 3 h of

exposure of Caco2 cells to both LR and DR MRPs treatments (Figure 3.17). There was a

trend for a decrease in Caco2 SOD activity when cells were treated with LR and DR

MRPs, and the effect was significant after 72 h of treatment (Figure 3.18).

124

Figure 3.16 Effect of AraSer MRPs extracts on glutathione (GSH) content in Caco2 cells. Cells were exposed to1.0 mg/ml green light roasted (LR) and dark roasted (DR) MRPs for 3 h (A), 24 h (B) and 72 h (C). Cells incubated in culture medium were used as control. Results are mean + SD, n=3. * indicates significant difference compared to corresponding control (p<0.05). Statistical analysis was done by oneway ANOVA with Dunnett’s multiple comparison tests.

125

Figure 3.17 Effect of AraSer MRPs extracts on glutathione peroxidase (GPX) activity in Caco2 cells. Cells were exposed to 1.0 mg/ml light roasted (LR) and dark roasted (DR) MRPs for 3 h (A), 24 h (B) and 72 h (C). Cells incubated in culture medium were used as control. Results are mean + SD, n=3. * indicates significant difference compared to corresponding control (p<0.05). Statistical analysis was done by oneway ANOVA with Dunnett’s multiple comparison tests.

126

Figure 3.18 Effect of AraSer MRPs extracts on superoxide dismutase (SOD) activity in Caco2 cells. Cells were exposed to1.0 mg/ml light roasted (LR) and dark roasted (DR) MRPs for 3 h (a), 24 h (b) and 72 h (c). Cells incubated in culture medium were used as control. Results are mean + SD, n=3. * indicates significant difference compared to corresponding control (p<0.05). Statistical analysis was done by oneway ANOVA with Dunnett’s multiple comparison tests.

127 3.3.11 Biological effects of Ara-Ser MRPs fractions on Caco-2 cells

3.3.11.1 Caco2 MTT response

Table 3.18 summarizes the MTT response of Caco2 cells exposed to 1.0 mg/ml crude

extract and Fractions II and III recovered from LR AraSer MR system, respectively.

Caco2 cells treated with MRPs crude extract for 3 h showed no significant changes in

cell viability. However, Caco2 cells incubated with Fraction III for 3 h produced a

significant (p<0.05) increase in cell viability, which contrasted the reduced cell viability

observed in Fraction II treated cells (p<0.05). Pretreatment of Caco2 cells with 1.0

mg/ml MRPs fraction III for 3 h prevented a H2O2 induced reduction in cell MTT

response, thus maintaining cell viability in a range that was common for nonstressed

control cells. A similar pretreatment of Caco2 cells with crude extract and Fractions II

failed to show protective effects. There was no significant change in Caco2 cell viability

after 24 h treatment of crude MRPs extract and Fractions II and III, respectively.

Moreover, the reduced cell viability induced by H 2O2 was not prevented by pretreatment

of cells with MRPs extract or related fractions for 24 h.

1 Table 3.18 Caco2 MTT response to MRPs with (+) and without () H 2O2 treatment MRPs 3h Treatment 24h Treatment H 2O2 + H 2O2 H 2O2 + H 2O2 Control 100+8 80+4* 100+6 80+4* Crude 103+2 73+3 88+2 71+4 Fraction II 82+4* 70+5 98+4 75+4 Fraction III 121+4* 101+2** 96+1 82+1 1 Data (% control) represent means + SD obtained for triplicate samples measured in triplicate. Cells treated with MRPs samples alone were compared to negative control ( H2O2), and cells pretreated with MRPs samples and further challenged by H2O2 were

compared to positive control (+H2O2). * indicates significant difference in comparison ** with negative control (H2O2), and indicates significant difference in comparison with positive control (+H2O2). Statistical analyses were done by oneway ANOVA with Dunnett’s tests (p<0.05).

128 3.3.11.2 Caco2 cellular glutathione content

Treating Caco2 cells with 5 mM H 2O2 for 2 h reduced the cellular GSH content by 16%,

which was significantly lower (p<0.05) from the nonstressed, negative control. The

exposure of Caco2 cells to the crude extract, as well as to Fractions II and III, derived

from the LR AraSer MR system, also significantly (p<0.05) reduced the cellular GSH

content (Figure 3.19). There was no significant difference in GSH content between cells pretreated with MRPs and the positive control (+H2O2).

Figure 3.19 Effect of MRPs extracts and associated fractions derived from light roasted (LR) AraSer MR system on Caco2 cellular glutathione (GSH) contents after 24h of treatment. Without () H2O2: ; with (+) H2O2 : . GSH content in cells that were treated with MRPs samples alone was compared to that of negative control (H2O2), and GSH content in cells that were pretreated with MRPs samples and further challenged by

H2O2 was compared to that of control (+H2O2). * indicated significant difference in comparison with corresponding control (p<0.05, oneway ANOVA with Dunnett’s tests).

3.3.11.3 Caco2 antioxidant enzyme activities

The antioxidant enzyme activities of Caco2 cells after a 24 h treatment with MRPs crude extract and related Fractions derived from LR AraSer model system are presented in

Figure 3.20. Moreover, the same treatments with added H 2O2induced stress are also

shown in Figure 3.20. No significant change in GPX activity was observed in Caco2

129 cells treated with AraSer MRPs crude extract, or when exposed to Fraction III recovered from the AraSer MR model system. In contrast, treatment of cells with Fraction II from

the same crude extract resulted in a significant reduction in GPX activity (p<0.05).

Treatment of cells with H2O2 alone also produced a significant increase (p<0.05) in Caco

2 GPX activity compared to the negative control (H2O2). Cells pretreated with Fraction

II recovered from LR AraSer MR system exhibited reduced (p<0.05) GPX activity after

H2O2 treatment, compared to the positive control (+H2O2).

Treating Caco2 cells with H2O2 did not result in significant changes in cellular SOD activity. However, Caco2 cells pretreated with Fraction II recovered from AraSer MRPs extract showed significantly reduced SOD activity after further H2O2 exposure (p<0.05).

A significant (p<0.05) increase in CAT activity occurred in Caco2 cells treated with

H2O2. The activity of CAT was not significantly altered when cells were exposed to either the crude extract or the ultrafiltration fractions recovered from AraSer model system, respectively. Pretreatment of Caco2 cells with MRPs crude extract did not prevent the increase in CAT activity caused by a H2O2 assault. However, Caco2 cells pretreated with Fractions II and III recovered from MRPs crude extract had lower

(p<0.05) CAT activity compared to the positive control (+H2O2).

No significant change in Caco2 GR activity was found between treatments and control cells (data in Appendix Table 12).

130

Figure 3.20 Effect of light roasted (LR) AraSer MRP extracts and fractions on Caco2 cellular antioxidant enzyme activities after 24h of treatment. A: glutathione peroxidase (GPX); B: superoxide dismutase (SOD); C: catalase (CAT). Without () H2O2: ; with (+) H2O2 : . Cells treated with MRPs samples alone were compared to negative control (H2O2), and cells pretreated with MRPs samples and further challenged by H2O2 were compared to positive control (+H2O2). * indicates significant difference in comparison ** with negative control (H2O2), and indicates significant difference in comparison with positive control (+H2O2). Statistical analyses were done by oneway ANOVA with Dunnett’s tests (p<0.05).

131 3.3.12 Gene regulation of MRPs on the human oxidative stress and antioxidant defense system (HOSAD) in Caco-2 cells

Results demonstrating the potential effect of MRPs derived from AraSer model systems

on Caco2 antioxidant gene regulation are presented in Table 3.19. After treatment of

cells with the LR AraSer MRPs, significant changes in the expression of 10 genes

associated with the HOSAD were observed using a PCR array (fold change>2, p<0.05).

Of particular interest was the finding that expression of Caco2 GPX2 , PRDX4 was up

regulated 4.57fold and 2.36fold, respectively, whereas expression of iNOS was up

regulated 2.90fold. The expression of PRG3 was downregulated 8.21fold, and SEPP1

was downregulated 3.11fold. In contrast, Caco2 cells treated with the MRPs derived

from DR AraSer model system had 5 genes that were upregulated and no down

regulation of genes. The expression of several genes that was altered in Caco2 cells after

treatment with coffee samples was also markedly changed in MRPs treated cells (Figure

3.21).

132 Table 3.19 Genes differently expressed in Caco2 after incubation with AraSer MRPs 1 Gene symbol Treatment LR DR GPX2 +4.57 +8.79 SELS +2.41 n/c PRDX4 +2.36 n/c iNOS +2.09 +2.14 MPO +2.05 n/c PRG3 8.21 n/c PTGS1 4.28 n/c SEPP1 3.11 n/c PRNP 2.07 n/c MSRA 2.01 n/c ALOX12 n/c +4.42 ANGPTL7 n/c +4.05 CCS n/c +2.14 1 Data represent fold changes of differently expressed genes in treated cells compared to control cells (n=3, fold change>2, p<0.05). “+” = upregulation; “” = downregulation; n/c = no significant change or fold change<2. The name and function of these genes are described in Appendix (Table 8).

133

Figure 3.21 Antioxidant genes expression in Caco2 cells treated with light roasted (LR: ), dark roasted (DR: ) AraSer MRPs compared to those in control cells. A: up regulated gene; B: downregulated gene. ** indicates p<0.01, * indicates p<0.05. GPX2 = glutathione peroxidase 2; SELS = Selenoprotein S; PRDX4 = peroxiredoxin 4; iNOS = inducible nitric oxide synthase; PRG3 = proteoglycan 3; PTGS1 = protaglandin endoperoxide synthase 1; SEPP1 = selenoprotein P; MSRA = methionine sulfoxide reductase A. The name and function of these genes are described in Appendix (Table 8).

134 3.4 Discussion

3.4.1 Chemical characteristics of coffee extracts and model MRPs

The many challenges associated with the isolation and identification of individual products of the Maillard reaction derived from coffee has prevented a clear understanding of the structureactivity relationships between coffee MRPs and antioxidant and bioactive potential, respectively. In Experiment IIa of the present study, nonroasted, green coffee beans were used as a negative control with the understanding that this food matrix, albeit

containing bioactive phenolics, such as CGA and caffeine, did not contain products of the

Maillard Reaction. Model MRPs recovered from heated AraSer reactants were used in

Experiment IIb as a reference for characterizing the bioactivity of MRPs generated during

coffee roasting.

The formation of MRPs and the degradation of phenolics in roasted coffee beans were

characterized by a battery of fluorescence, UVvis spectra and tristimulus color

measurements made on nonroasted and roasted coffee beans, as well as the model MRPs.

From the fluorescence spectra, it could be ascertained that phenolics, particularly CGA

were involved in the formation of both high molecular weight and low molecular weight

MRPs during coffee roasting process. This finding agrees with results obtained in the previous chapter and also from a previous study, that showed CGA present in GB were

incorporated into the high molecular weight melanoidin polymers (MW>10KDa)

(Bekedam et al. , 2008d). It is also likely that high molecular weight phenolics, such as

tannins (Savolainen, 1992) and phenolic glycosides (Clifford, 1985), were retained in the

coffee fraction that contained the high molecular weight melanoidins (MW>10KDa).

However, there is no indication that these polyphenolic compounds were incorporated

135 into the melanoidin structures (Clifford, 1985; Savolainen, 1992). Since no correlation

could be made between the changes in fluorescence and molecular weight of derived

model MRPs after roasting, it remains uncertain if the generation of high molecular

weight MRPs can be explained simply by the polymerization reactivity of low molecular

weight MRPs.

Making a comparison between the UVvis spectra of the high molecular weight fractions

derived from the AraSer model system and coffee extracts, respectively, showed some

degree of similarity in the chromophoric substructures that could be attributed to melanoidin formation in both the model, and the food system. A previous study has reported similar UVvis spectra between melanoidins derived from glucosecasein model system and those isolated from dark beer (Hofmann, 1998b). The results therefore support, and extend the hypothesis that melanoidins derived from sugaramino acid, or sugarprotein MR model systems, can be used to model the more complex food systems that contain many more discrete chromophores (Figure 4.1) (Hofmann, 1998b). The basic melanoidin skeletons from different MR sources will likely depend on the type of sugar, amino acids and proteins that are involved in the reaction. However, the results herein indicate that some similarities in the chromophoric substructures or chromophoric low molecular MRPs indeed exist, and thus confirm the crosslinking expected to occur between amino acid residues that form the high molecular weight melanoidins.

136

Figure 4.1 UVvis spectrum of a typical melanoidin ( ) and of individual chromophoric substructures () (Hofmann, 1998b).

The Hunter L a b color scale and the absorbance readings taken at 420 nm have also been used previously to monitor the development of colored compounds during Maillard reaction (Kitts and Hu, 2005; Gokmen and Senyuva, 2006). However, these methods were found to be insensitive in this study for showing detectable colorimetric differences between LR and DR in both coffee extracts and the model MRPs. The results however, do show that there was a positive correlation between the darkness and browning, with increased molecular weight of the constituents recovered in the individual fractions. The degree of browning was found to be related more to the type of sugar than the degree of roasting, which agrees with other findings from different thermally processed food matrices, or model MRPs (Kurosaki et al. , 1989; Lingnert, 1990; Jing and Kitts, 2002).

137 3.4.2 Antioxidant activity and reducing power of coffee constituents

3.4.2.1 The underlying antioxidant mechanisms of MRPs derived from coffee extracts

It has been shown in the previous studies that coffee brews exhibit antioxidant activity in both the ORAC and TEAC assays (Pellegrini et al. , 2003; del Castillo et al. , 2005;

DelgadoAndrade et al. , 2005; SanchezGonzalez et al. , 2005), which agrees with the findings in the present study. However, the concentration of phenolic nonMaillard antioxidants in coffee brews was relatively high in some studies due to the mild roasting condition employed. In other studies, the phenolic content or the roasting conditions were not provided and therefore results of antioxidant activity could not be explained completely. It is therefore difficult to gain insight into the antioxidant property associated with coffee MRPs due to interference of phenolics and the complicity of the reaction products. The roasting conditions used for the commercial LR and DR coffee beans in the present study were more severe compared to those used in other studies (del Castillo et al. , 2002; del Castillo et al. , 2005; Sacchetti et al. , 2009). In this study, the plant phenolic antioxidants were mostly degraded during the roasting process, which was demonstrated from the CGA absorbance spectra derived from the coffee extracts. Other workers have also reported 84 to 95% loss of CGA in coffee extracts prepared using the same coffee bean species and similar roasting conditions and brewing methods (Trugo, 1984; Budryn et al. , 2009). Therefore, with this in mind, it is most likely that the antioxidant activity of the coffee extracts and related fractions reported in the present study are depended mostly on the presence of coffee MRPs. Furthermore, the present results also demonstrate that

MRPs derived from coffee extracts exhibit peroxyl radicals scavenging activity in the

ORAC assay by a transfer of hydrogen atoms to oxygen radicals, also referred to as the

138 HAT (hydrogen atom transfer) antioxidant reaction mechanism (Mayer, 2004). Similar

antioxidant capacity of coffee MRPs also involves the transfer of single electron (SET mechanism) to oxidants, which was shown herein by the TEAC and Gu’s RP assays,

which reduces the ABTS radicals and the ferric ions, respectively (Mayer, 2004).

3.4.2.2 Antioxidant activity of phenolic constituents in green coffee extracts and MRPs

derived from roasted coffee extracts

Despite several reports on the antioxidant activity of the roasted coffee brews/extracts,

the antioxidant capacity of coffee extracts derived from green, unroasted coffee beans, in particular, has not been well studied. Results from the present study showed that GB

extracts had higher antioxidant activity compared to roasted coffee extracts, which agrees

with previous findings (Daglia et al. , 2000; GomezRuiz et al. , 2008). Chlorogenic acids

monoester, 5CQA, was suggested to possess the highest antioxidant activity among the

different CGA esters present in GB in scavenging hydroxyl and peroxyl radicals (Daglia

et al. , 2000; Daglia et al. , 2004). The results from the present study indicate, however,

that CGA diesters and mixed esters have higher peroxyl radical scavenging activity

compared to CGA monoesters (CGA diesters mostly present in Fraction III of GB extract

and monoester mostly present in Fraction IV of GB extract). There are few studies that

have reported on the antioxidant activity of CGA diesters and mixed esters, albeit the

chemical structures are well known (Figure 4.2) (Farah and Donangelo, 2006).

139 A

B

C

D

Figure 4.2 Chlorogenic acids and related compounds according to chemical characteristics. (A) Basic compounds: CA: caffeic acid; FA: ferulic acid; pCoA: p coumaric acid; (B) monoesters of quinic acid with hydroxycinnamic acids: caffeoylquinic acids (CQA), with 3 isomers (3, 4 and 5CQA); (C) diesters of quinic acid with caffeic acid: dicaffeoylquinic acids (diCQA), with 3 isomers (3,4diCQA; 3,5 diCQA; 4,5diCQA); and (D) mixed esters: six mixed diesters of caffeoylferuloylquinic acids (CFAQ). (Farah and Donangelo, 2006)

The antioxidant activity of MRPs derived from roasted coffee extracts is relatively lower

compared to that of the CGA present in the GB extract, based on the data from the

current study. This result extents the finding from a previous study, where a lower

antioxidant activity of coffee melanoidins compared to CGA was reported (Delgado

140 Andrade et al. , 2005). Nevertheless, the antioxidant activity of coffee extract is still higher than that of green tea and black tea extracts (Cao et al. , 1996), and coffee MRPs have been mentioned to be great contributors to the total dietary antioxidant intake

(Svilaas et al. , 2004).

In agreement with the results reported in Chapter II and also a previous study (Somoza et al. , 2003), results herein showed that low molecular weight (MW<1KDa) MRPs derived from roasted coffee contribute to the most of the total ORAC, and TEAC antioxidant activity as well as the reducing power of the whole coffee extracts. This result is partially due to a greater proportion of compounds present in the low molecular weight fractions that exhibit antioxidant capacity, compared to compounds that make up the high molecular weight fractions (MW>1KDa) recovered from the coffee extracts. Our results also indicate that no positive or negative correlation exists between the antioxidant activity of coffee MRPs and the molecular weight of the compounds in these fractions. In addition, coffee MRPs showing the highest antioxidant activity in scavenging peroxyl radicals, expressed by an underlying HAT mechanism, did not have the highest ABTS radicals scavenging activity which occurs by a SET mechanism. In a recent study, Gu et al. , reported that MRPs of different molecular weights derived from caseinglucose MR model system exhibited distinctly different antioxidant activities that included scavenging free radicals, sequestering metal ions and inhibiting lipid oxidation (Gu et al. , 2009).

These results indicate that MRPs could act simultaneously with different mechanisms of action.

141 3.4.2.3 ORAC, TEAC and Gu’s RP assays assessing the antioxidant activity of coffee

As mentioned, the ORAC assay measures the antioxidant activities of chainbreaking

antioxidants against peroxyl radicals, based on the HAT mechanism (Cao et al. , 1993; Ou

et al. , 2001; Mayer, 2004). Compared to the ORAC assay used in other studies that test the sample of a single concentration (Cao et al. 1993; Ou et al. 2001), the improved

ORAC assay by Kitts and Hu measures the peroxyl radical quenching capacity of an antioxidant over a range of different concentrations and calculates the ORAC value by dividing the slope of the regression equation of the sample by the slope of Trolox calibrator (Kitts and Hu, 2005; Saenz et al. , 2009). This modification offers a greater

sensitivity and precision of measurement and produces consistent results for assessing the peroxyl radical scavenging activity of coffee extracts in the present study. Different

Trolox equivalents (TE) units have been used to express the ORAC value of coffee,

including TE/liter of coffee brew, TE/gram of freezedried coffee extract, and TE/gram of

coffee beans. These differences make the comparison of ORAC antioxidant activity of

coffee brews/extracts between studies difficult. The use of Trolox equivalents/ gram of

coffee extract (dry matter basis) represents a measure of antioxidant activity from coffee

constituents that is independent of the difference in the concentration of coffee brews/extracts, where there is the use of a Trolox calibrator.

The measurement of the SET specific antioxidant action, which also reflects the biological function of the antioxidant is relatively difficult to achieve compared to assessing the HAT mechanism (Ou et al. , 2002). With the TEAC assay, measuring the

ABTS radical scavenging activity, and FRAP assay (ferric reducing antioxidant power), or Gu’s RP assay that measures the total reducing power of the antioxidant, simple and

142 very common used methods were popular for assessing antioxidant activity of dietary

components (Karakaya et al. , 2001; Pellegrini et al. , 2003). The results of the current

study indicate that Gu’s RP assay is a relatively insensitive method compared to the

TEAC assay in assessing the SET activity of MRPs derived from the coffee extract. In

addition, reducing power measurement estimates only the ferric reducing activity, and is

not necessarily relevant to the total antioxidant capacity. Therefore, we suggest that

ORAC and TEAC assays are preferable methods to evaluate the antioxidant activity of

MRPs derived from coffee extract. Additional assays that have been used to measure

• antioxidant capacity against other ROS, such as O 2 , ONOO , and singlet oxygen, have not been employed to measure the total antioxidant activity of coffee and MRPs derived from coffee extracts against multiple sources of ROS.

3.4.2 Biological effects of green coffee, roasted coffee and model MRPs on the antioxidant defense system in Caco-2 cells

3.4.2.1 MRPs derived from coffee and MR model systems inhibit the activity of antioxidant enzymes in Caco2 cells

Despite the strong in vitro antioxidant activity of MRPs derived from either complex food systems or simple model reaction systems, supporting data on the antioxidant effect of MRPs from in vivo studies are mostly inconsistent. Feeding rats a MRPcontaining diet has been shown to increase the total antioxidant capacity in plasma (Somoza et al. , 2005).

Slightly increased plasma content of GSH was reported in humans that consumed five cups of coffee per day for a week (Esposito et al. , 2003). In contrast, a recent study has reported no changes in serum biomarkers of oxidative stress and antioxidant defenses (e.g.

143 hydroperoxides content, CAT, SOD and GPX) after coffee consumption (Seiquer et al. ,

2008). In another study, an increased oxidative stress associated with coffee consumption

has been reported (Sakamoto et al. , 2003). It is safe to conclude that the in vivo

antioxidant mechanisms associated with MRPs are not fully understood, nor have they been demonstrated in human subjects.

In the present study, treating Caco2 cells with MRPs extracts derived from coffee and

MR model systems resulted in a reduced cellular GSH content that corresponded to

reduced activities of specific antioxidant enzymes. The only exception of this was an

increased cellular GPX activity after 24 h coffee incubation. A similar result has been

reported in a previous study that incubated HepG2 cells with coffee melanoidins showing

a reduced cellular GSH content (Goya et al. , 2007). In addition, treatment of human

lymphocytes and Int407 cells with MRPs derived from different MR model systems has been reported to result in a reduced cellular GSH content and lower CAT, GR, GPX, and

SOD activities (Yen et al. , 2002; Jing and Kitts, 2004b).

At present, the mechanism for the inhibitory effect of MRPs on the activity of specific

antioxidant enzymes remains unknown; however, several hypotheses have been

suggested. Antioxidant enzymes require metal cofactors in order to reach maximum

efficiency. For example, SOD consists of proteins cofactored with copper, zinc, or

manganese (Harris, 1992). Iron is also required as a for CAT (Tanswell and

Freeman, 1984; Minotti and Aust, 1992). The metal sequestering activity of melanoidins

has been shown for copper (Wijewickreme et al. , 1997; Wijewickreme and Kitts, 1998a)

and iron (Morales et al. , 2005), two cofactors for SOD and CAT activity, respectively. It

is therefore plausible that part of the inhibitory effect of MRPs, particularly the high

144 molecular weight melanoidins, could be related to metal sequestering activities

(Wijewickreme et al. , 1997; Wijewickreme and Kitts, 1998a, b; Morales et al. , 2005).

However, the mechanism for the enzyme inhibitory action associated with low molecular

weight MRPs appears to be different from melanoidins, since the metalchelating properties of low molecular weight MRPs are limited (DelgadoAndrade and Morales,

2005). It is also known that enzymes containing a functional group, such as the heme for CAT, require a catalytic threedimensional form in order to be reactive

(Putnam et al. , 2000). Binding to these enzymes in either the active site or the nonactive site can reduce the activity to react with the substrate (Wyvratt and Patchett, 1985;

Putnam et al. , 2000). A recent study has reported that methylglyoxal, a product of MR,

can exhibit an inhibitory effect on GPX activity by binding to the GSH and

inactivate GPX in a time and dosedependent manner (Park et al. , 2003). Other α

dicarbonyl MRPs, such as 3deoxyglucosone, glyoxal and phenylglyoxal have also been

reported to inactivate GPX (Park et al. , 2003). In addition, methylglyoxal has been found

to inhibit the activity of SOD, GST, and CAT, and reduce GSH content in both in vitro

and in vivo studies (Choudhary et al. , 1997). These particular αdicarbonyls were

recovered in both coffee and model MRPs. Thus, the potential of αdicarbonyl MRPs to

act as enzyme inhibitors, competing with the substrate for the active center of the specific

antioxidant enzymes, and lowering the antioxidant enzyme activity, could explain the

similar reduced enzyme activities for GPX, CAT and SOD observed for both coffee and

model derived MRPs.

145 3.4.2.2 Effect of melanoidins on cellular oxidative stress induced by different oxidants

In the present study, pretreatment of Caco2 cells with MRP extract and the high

molecular weight melanoidin fraction, respectively, derived from coffee and model systems showed no protection against H 2O2induced cellular oxidative stress. Similar

findings have been reported with model MRPs having no protection on DNA damage in

human lymphocytes induced by H 2O2 (Yen et al. , 2002). In contrast, one study reported a protective effect of coffee melanoidins against tert butylhydroperoxide ( tBOOH)

induced damage in HepG2 cells (Goya et al. , 2007). These results indicate that the protective effect of MRPs could be dependent on the source of the prooxidant that was used to induce oxidative cell damage. tBOOH , a well know oxidizing agent, leads to a loss of CAT activity and a subsequent increase in GPX and GR activity that reflects cell compensation to the loss of CAT against hydrogen peroxide and other peroxides

(Pichorner et al. , 1993; Goya et al. , 2007). In the presence of a HAT agent, tBOOH is converted to tbutanol, which is a nonreactive substrate (Pichorner et al. , 1993). As discussed in the previous section, MRPs possess strong antioxidant activity by donating hydrogen atoms (HAT mechanism), which can explain the protective effect against t

BOOHinduced changes in the cellular antioxidant defense reported by Goya et al.

(2007). This result contrasts the findings of the present study that showed no affinity for

MRPs to scavenge H 2O2. In fact, several researches have also reported MRPs derived

from both Maillard model systems and food systems (e.g. coffee) actually produce H 2O2

as a byproduct of the reaction (Namiki and Hayashi, 1975; Nagao et al. , 1986; Roberts

and Lloyd, 1997; Yen et al. , 2002; Muscat et al. , 2007; Hegele et al. , 2009). It is

important to also note that H 2O2 production from MRPs has been reported to increase in

146 the cell culture medium and result in cytotoxicity of HepG2 cells (Hegele et al. 2009). In

the present study, exposing Caco2 cells to H 2O2, following a pretreatment with MRPs provided no protection and moreover actually increased cytotoxicity. This result may be

explained by the fact that the amount of H 2O2 generated from the presence of MRPs in culture medium exposed to cells, along with the actual addition of H 2O2, collectively contributed to the observed Caco2 cytotoxicity. The results of this study also showed that a greater cytotoxic response to MRPs was attributed to both the crude extract and the high molecular weight (MW>1KDa) fraction, but not to the low molecular weight fractions. Therefore, it is possible that this difference between high molecular weight and low molecular weight MRPs reflects that affinity of high molecular weight melanoidins to generate H 2O2.

3.4.3 Coffee and the expression of Redox-sensitive genes in Caco-2 cells

3.4.3.1 Coffee constituents and regulation of antioxidant gene expression in Caco2 cells

Data from the present study indicate that coffee extracts contain potential signaling molecules that can both up and downregulate the expression of specific genes involved in the oxidative stress and antioxidant defense systems in Caco2 cells. Extracts derived from nonroasted coffee beans showed little effect on the antioxidant gene expression in

Caco2 cells, which indicates that caffeine and natural phenolics are unlikely to be the main signaling molecules. Treating Caco2 cells with coffee extracts resulted in similar changes in the expression of specific antioxidant genes as treating Caco2 cells with H 2O2.

This observation indicates that H 2O2 is involved in the regulation of antioxidant gene expression by coffee. The finding of H 2O2 as a signaling molecule in coffeeinduced gene

147 regulation process also agrees with a previous study, which showed that coffee induced

the activation and nuclear translocation of the transcription factor NFкB (nuclear factor

kappa B) via the generation of H 2O2 (Muscat et al. , 2007). In addition, MRPs may also be involved in Caco2 cell gene regulation after coffee treatment, as demonstrated by the result that the expression of some genes (e.g. MBL2 , SELS , PRG3 , and MSRA ) was changed in coffee and model MRPs treated cells, but not in H 2O2 treated cells. The potential of coffee to regulate antioxidant gene expression has been reported in previous

studies, where coffee induced the expression of genes that contain antioxidant response

elements (ARE) (Cavin et al. , 2008; Higgins et al. , 2008). Coffee diterpenes, cafestol and

kahweol (C+K), were reported to be responsible for the specific

antioxidant/chemopreventive gene regulatory effects associated with coffee (Cavin et al. ,

2002; Higgins et al. , 2008). However, it is important to remark that significant levels of

C+K are only found in some special coffee preparations, such as Turkish and boiled

coffees (Gross et al. , 1997). In most coffee brews, for example, the filtered coffee that

was used in the present study, the concentration of diterpenes would not be an issue and

therefore not attributed to a biological response (Gross et al. , 1997; Cavin et al. , 2008).

In general, data from the current study indicates that specific MRPs present in the

different coffee extracts, in addition to the possible generation of the H 2O2 in the cell culture system, collectively produced the genespecific effects observed in the PCR array.

This is the first time the expression of a panel of antioxidant genes in a cell line exposed to coffee has been reported.

148 3.4.3.2 Coffee induced the expression of specific AREdriven antioxidant genes

In the present study, coffee induced the expression of GPX2 (gastrointestinal glutathione peroxidase), SRXN1 (sulfiredoxin 1), TXNRD1 (thioredoxin reductase 1), PRDX1

(peroxiredoxin 1), PRDX4 (peroxiredoxin 4) and PRDX6 (peroxiredoxin 6) in Caco2 cells. These genes all contain antioxidant response elements (ARE) in the promoters, which are sensitive to oxidative stress (Lee et al. , 2003; Singh et al. , 2006; Soriano et al. ,

2008; Chowdhury et al. , 2009). In contrast, the expression of CAT (catalase), which also contains ARE sequence in the promoter, was downregulated in Caco2 cells after coffee treatment. This result indicates that the induction of AREdriven genes in Caco2 cells by coffee is likely selective and specific. It has been reported that catalase inhibits the activation of AREdriven genes (Lee et al. , 2001; Holland et al. , 2009). Therefore, it is possible that the downregulation of CAT expression in Caco2 cells by coffee may facilitate the induction of specific AREdriven genes. The induction of other AREdriven genes by coffee has been reported previously, which include the genes for NQO1,

UGT1A6, HO1, AKR7A1, GSTP1, GSTA1, GSTA3, GSTA4, GSTA5, and GCLC

(Cavin et al. , 2008; Higgins et al. , 2008). The induction of AREdriven genes by coffee was also shown to be concentrationdependent and cell typespecific (Higgins et al. ,

2008). It has been suggested that altering the expression of AREdriven genes is a key

molecular mechanism for detoxification and chemopreventive effects (Hayes et al. , 1999).

For example, glutathione peroxidase and peroxiredoxin are important for cellular defense

against H 2O2 (Sies, 1993; Rhee et al. , 2005) . These peroxidases use GSH and/or

thioredoxin as electron donor for detoxification of H 2O2 (Sies, 1993; Fisher et al. , 1999;

Manevich and Fisher, 2005). It is also well known that glutathione peroxidase is the

149 primary defense against H 2O2 at high concentrations (Masaki et al. , 1998; Wijeratne et al. ,

2005). Oxidized GSH and thioredoxin are converted to reduced forms by glutathione

reductase and thioredoxin reductase (TXNRD1). In the present study, the expression of

TXNRD1 was induced in Caco2 cells after coffee treatment. However, the mRNAs of

glutathione reductase and glutathione synthetase were not induced by coffee. This

specific induction of glutathione peroxidase and peroxiredoxin, but not glutathione

reductase and glutathione synthetase in Caco2 cells after coffee treatment, corresponds

to the reduced cellular glutathione content observed in Caco2 cells after coffee treatment.

These results indicate that coffee treatment of cells could induce oxidative stress, which

leads to the induction of specific antioxidant genes. The overall effect of coffee treatment

may be an increased resistance to oxidative damage by maintaining a certain level of

oxidative stress in cells. In fact, it has been suggested in other studies that the molecular

mechanism of an antioxidant action involves the generation of ROS as second

messengers for antioxidant gene induction (Favreau and Pickett, 1991; Li and Jaiswal,

1994; Pinkus et al. , 1996; Yang et al. , 2006; Holland et al. , 2009).

The transcription factor Nrf2 (Nuclear factorerythroid 2 p45 subunitrelated factor 2) has been implicated to be the central protein that interacts with the ARE to activate ARE

driven gene transcription constitutively, or in response to an oxidative stress signal

(Nguyen et al. , 2003; Jaiswal, 2004). Data from the present study indicate that coffee

induced oxidative stress, particularly the generation of H 2O2 in the cell culture system

(Hegele et al. , 2009), may function as a signal that activates Nrf2dependent induction of

AREdriven gene expression. Indeed, H 2O2 treatment of rat cardiomyocytes has been

reported to cause a rapid increase in the translation of Nrf2 protein (PurdomDickinson et

150 al. , 2007). The regulation of Nrf2/ARE signaling pathway by coffee constituents, C+K

and caffeine, has been addressed in previous studies (Cavin et al. , 2002; Okano et al. ,

2008). C+K disrupt the cytoplasmic Keap1Nrf2 complex through thiol modification of

cysteine residues in Keap1, therefore releasing Nrf2 and permitting its translocation to

nucleus, where Nrf2 transcriptionally activates AREdriven genes (DinkovaKostova et

al. , 2002). Caffeine activates MAPK/ERK signal pathway so as to phosphorylate Nrf2

and permits its translocation to the nucleus, where Nrf2 activates AREdriven genes

(Okano et al. , 2008). Taken together, these reported results indicate that coffee extracts

may induce Nrf2/ARE signaling pathway by increasing the translation and nuclear

translocation of Nrf2, which involves different coffee constituents.

3.4.3.3 Coffee and iNOS induction

Coffee treatment of Caco2 cells induced the mRNA of iNOS (inducible nitric oxide

synthase), which produces nitric oxide, a major endogenous modulator with diverse biological actions (Nussler and Billiar, 1993; YetikAnacak and Catravas, 2006; Chen and Kitts, 2008a). The expression of iNOS in Caco2 cells can be induced by certain cytokines, such as interleukin1 (IL1) and tumor necrosis factor (TNF) (Lowenstein et al. , 1992). A recent study demonstrated that oxidative stress can also lead to the up regulation of iNOS as well as eNOS (endothelial nitric oxide synthase) in vivo (Zhen et al. ,

2008). The proposed mechanism of oxidative stressinduced upregulation of iNOS is mainly through the activation of redoxsensitive transcription factor, NFкB (Xie et al. ,

1994; Liu et al. , 1997; Zhen et al. , 2008). Notably, it has been reported that coffee and model MRPs extracts led to nuclear translocation of NFкB in macrophages via the generation of H 2O2 (Muscat et al. , 2007). Hydrogen peroxide has been suggested to be

151 the primary activator of NFкB signaling pathway (Muscat et al. , 2007; Zhen et al. , 2008).

In addition to one NFкB binding site, the promoter region of iNOS gene contains two binding sites for activator protein (AP)1, which is also important for the transcriptional

induction of iNOS (Chu et al. , 1998). It has been reported in a previous study that caffeine induced the expression of AP1 protein (Svenningsson et al. , 1995). Thus, these data suggest that coffee induce the expression of iNOS through activation of AP1 and

NFкB. Caffeine may augment or synergize H 2O2induced expression of iNOS , which is

supported by the observation of greater induction of iNOS in coffee treated cells compare to cells treated with H 2O2 and model MRPs, respectively.

3.4.3.4 Proposed pathway of hydrogen peroxide generation in coffee and Maillard

reaction systems

The potential of coffee and MRPs to generate H 2O2, especially in the cell culture system, is intriguing. High levels of hydrogen peroxide can cause oxidative stress, which is an important cause of cell damage. However, it has also been suggested that H 2O2 triggers different signaling pathways in a concentration dependent manner and this results in redoxmediated changes in the function of enzymes, which include antioxidant enzymes and other proteins that protect cells against injury from ROS (Babior, 1997; Sablina et al. ,

2005; Bossis and Melchior, 2006). Based on the current understanding of the schemes of the Maillard reaction and the autoxidation of monosaccharide and polyphenols (Wolff and Dean, 1987; Elgawish et al. , 1996; Mochizuki et al. , 2002), the mechanism of how

• ROS, including O 2 and H 2O2 are formed in coffee brews is proposed here for the first time (Figure 4.3). The general hypothesis is that many coffee constituents contain a special enediol group or an enediollike chemical group (Figure 4.3A), which can

152 • generate superoxide anions (O 2 ) when in the presence of trace metal catalysis (Figure

4.3, Scheme I). The basic scheme I is based on the available mechanisms of the formation

of superoxide anions from Amadori products, hydroxyaldehyde, and polyphenols, which

are all coffee components (Wolff and Dean, 1987; Elgawish et al. , 1996; Yaylayan et al. ,

1998; Nakanishi et al. , 2001; Akagawa et al. , 2003). The initiation of Scheme I reaction

is a metalcatalysed process (Wolff and Dean 1987; Elgawish et al. 1996; Akagawa et al.

2003), which forms a very unstable resonance structure containing a delocalized electron

system. This electron when transferred to an oxygen molecule will form a superoxide

anion. Superoxide anions generated are spontaneously converted to H 2O2 through

dismutation reaction according to Scheme II (Figure 4.3) (Babior, 1997; Akagawa et al. ,

2003).

Data from the present study and former studies indicate that hydrogen peroxide formation

is associated with melanoidins. Previous studies have demonstrated that free radicals,

including superoxide radicals are generated through sugar fragmentation and Amadori

rearrangment during the early stage of the Maillard reaction (Namiki and Hayashi, 1983;

Kawakishi et al. , 1994; Roberts and Lloyd, 1997). Both sugar fragments and Amadori products have been identified in the melanoidins structures (Cämmerer and Kroh, 1995).

Phenolics have also been detected in melanoidins derived from coffee brews (Bekedam et

al. , 2008d). These structural components of melanoidins all in common have the affinity

to generate superoxide radicals and H 2O2, especially in phosphate buffer, which provides

the optimal pH and also contains trace metal ions (Thornalley et al. , 1984; Kawakishi et al. , 1994; Akagawa et al. , 2003). The fact that melanoidins chelate metal ions may also favor the structure components generating H 2O2.

153 A. Enediol group OH OH

R1 CC R2

Scheme I: the basic scheme

- OH OH OH O O O O O n+ (n1)+ M M O2 O2 R1 CC R2 R1 CC R2 R1 CC R2 R1 CC R 2

i. Amadori product CH NH R2 CH NH R2 2 2 CH2NH2R2 2 2 CH2NHR2 n n+ M(n1)+ O O o C OH C OH M C O 2 2 ati C O iz ol en C OH C O 3 C O C O 2, CH2NH2R2 R1 R1 R1 R1 C O

HO C H 1, 2 R1 en o HC NHR2 HC HC NR2 HC NR2 liz NR2 at (n1)+ io Mn+ M O O2 n C OH C OH C O 2 C O

HO C H C O C O C O

R1 R1 R1 R1

ii. Hydroxyaldehyde

(n1)+ O O OH O - Mn+ M O O OH O O2 O2 R1 CC H R1 CC H R1 CC H R1 CC H dicarbonyl H

iii. Phenolics

OH OH n+ M(n1)+ O O M O2 O2

R OH R O R O R O

Scheme II

+ 2O2 + 2H H2O2 + O2 O H O + 2 + 2H2O 2 2 2OH

Figure 4.3 Proposed pathway of H 2O2 formation in coffee brews.

154

CHAPTER IV

GENERAL DISCUSSION AND CONCLUSIONS

155 4.1 GENERAL DISCUSSION

4.1.1 Chemical characteristics and antioxidant properties of coffee

The development of brown color during the roasting process of coffee beans is directly related to the Maillard reaction. The UV, fluorescence and color measurements were used to characterize and compare the chemical properties of MRPs derived from coffee and

MR model systems. A positive correlation between darkness and browning, with increased molecular weight of MRPs was observed. In contrast, the fluorescent property of MRPs did not show the same correlation with the molecular weight. No positive, or negative, correlation was observed between the antioxidant activities of MRPs and the degree of browning or the fluorescent intensity. Results from previous studies regarding the relationship between the antioxidant activity of MRPs and color and fluorescent properties are inconsistent (Monti et al. , 1999; Morales and JimenezPerez, 2001;

Murakami et al. , 2002; Chen and Kitts, 2008b). The discrepancies could due to different reactants, different reaction conditions, and different assays that were used to assess the antioxidant activity. However, taken together with the results from the present study, it can be conclude that color and fluorescent properties are useful indicators for the development of Maillard reaction, but are not directly correlated to the antioxidant potential of crude or fractionated MRPs. By extension, it can also be concluded that chromophore substructures of MRPs derived from coffee and model systems are not responsible for antioxidant action, which agrees with a previous finding (RufianHenares and Morales, 2007c).

156 4.1.2 Coffee, antioxidant enzymes and antioxidant genes

The results of the chemical antioxidant assays clearly showed that coffee has high

antioxidant activity by scavenging free radicals and also possesses reducing activity.

However, potential interaction of bioactive components in coffee and living cells resulted

in different antioxidant mechanisms of actions. In fact, a prooxidant property associated

with coffee was observed in Caco2 cells, which involved a reduced cellular glutathione

status after coffee treatment.

Experimental evidence indicates that antioxidant defense systems can act synergistically

or antagonistically, according to the level or activities of the systems involved (Chow,

1988). Coffee treatment of Caco2 cells enhanced the nonenzymatic antioxidant defense,

which may compensate for the negative effects on the activities of antioxidant enzymes.

In addition, hydrogen peroxide generated in the cell culture system as a result of coffee

exposure is capable of signaling and communicating critical information to activate

specific signaling pathways, which will influence various cellular processes that lead to

improved cell functions depending on the concentration of H 2O2 (Scandalios, 2005; Veal et al. , 2007). In the present study, the induction of AREdrive gene expression in Caco2 cells exposed to coffee extract indicated an enhanced cellular antioxidant defense system.

Results from the present study also indicate that coffee could regulate gene expressions in

Caco2 cells through Nrf2, NFκB and Ap1 signaling pathways. Many cell functions, such as inflammatory response and anticarcinogenic effects, are mediated either directly or indirectly through these same pathways (Tak and Firestein, 2001; Zhang and Gordon,

2004). In general, Caco2 cells adapted to the coffeeadded environment by altering the

157 expression of specific antioxidant and oxidative stress response genes, which are likely to result in a more responsive defense system when challenged by ROS.

158 4.2 CONCLUSION

Maillard reaction products were shown to be the prevailing antioxidants present in

roasted coffee extracts. Chromophore substructures of the coffee MRPs are not responsible for the antioxidant action. No positive, or negative, correlation exists between the antioxidant activity of coffee MRPs and molecular weight. The mechanisms of the antioxidant action associated with coffee MRPs involve both the hydrogen atom transfer

(HAT) mechanism and single electron transfer (SET) mechanism. Reducing power is relatively less important to predicting antioxidant capacity of MRPs.

Treatment of Caco2 cells with MRPs derived from both coffee extracts and model systems resulted in increased cellular oxidative stress, as evidenced by a reduced glutathione content that corresponded to modified activities of specific antioxidant enzymes, namely GPX, CAT and SOD. The inhibitory effect of coffee extracts on specific antioxidant enzymes is suggested to be related to the metal sequestering activity of high molecular weight MRPs and the presence of αdicarbonyl MRPs, which are potential enzyme inhibitors. In addition, exposing Caco2 cells to roasted coffee extracts resulted in many changes in the expression of genes that are involved in human oxidative stress and antioxidant defense system. The induction of AREdriven gene by coffee resulted in an increased endogenous defense mechanisms that could respond by enhanced protection against various types of environmental stresses. The induction of AREdriven genes by coffee is suggested herein to be associated with the Nrf2 signaling pathway. The specific induction of iNOS gene expression in Caco2 cells also indicates a possible role

for coffee intake in inflammatory and immune responses. Results indicated that the

159 activation of iNOS gene by coffee is likely to be mediated by the activation of the

transcription factors, AP1 and NFκB. Hydrogen peroxide generated in the cell culture

system as a consequence of coffee exposure, may serve as a signaling molecule that is

involved in the gene regulatory effect associated with coffee extracts. The generation of

H2O2 by coffee constituents was proposed. In addition, our data for the first time showed the antioxidant character of coffee MRPs on a molecular level. The overall gene regulatory effect of coffee constituents on Caco2 cells is more likely to be the combined effects of all bioactive constituents, which involves a complex interplay of chemical and biological reactions and responses.

160 4.3 SUGGESTIONS FOR FUTURE RESEARCH

(1) To investigate the specific effect of αdicarbonyl MRPs, for example, methylglyoxal

on the expression of antioxidant genes in Caco2 cells; the combined effect of

methylglyoxal and hydrogen peroxide on the activities of antioxidant enzymes and the

expression of antioxidant genes.

(2) To confirm the AREdriven gene activation by coffee in Caco2 cells on the protein level.

(3) To confirm the activation of Nrf2, AP1, and NFкB signaling pathways by coffee in

Caco2 cells; look at the effect of MRPs (high molecular weight and low molecular

weight) on the activation of these transcription factors.

(4) To investigate the effect of MRPs derived from coffee and model systems on the

cellular antioxidant defense system in fully differentiated Caco2 cells.

(5) To conduct in vivo study on the antioxidant effect of MRPs derived from coffee.

161 REFERENCES

Akagawa, M., Shigemitsu, T. & Suyama, K. (2003). Production of hydrogen peroxide by polyphenols and polyphenolrich beverages under quasiphysiological conditions. Bioscience Biotechnology and Biochemistry 67 (12): 26322640.

Ames, J. M. (1992). The Maillard reaction. Biochemistry of food proteins 4: 99153.

Ames, J. M., Bailey, R. G. & Mann, J. (1999a). Analysis of furanone, pyranone, and new heterocyclic colored compounds from sugar glycine model Maillard systems. Journal of Agricultural and Food Chemistry 47 (2): 438443.

Ames, J. M., Wynne, A., Hofmann, A., Plos, S. & Gibson, G. R. (1999b). The effect of a model melanoidin mixture on faecal bacterial populations in vitro. British Journal of Nutrition 82 (6): 489495.

Anderson, M. E. (1985). Glutathione and glutathione disulfide in biological samples Methods in Enzymology 113 : 548555.

Anese, M. & Nicoli, M. C. (2003). Antioxidant properties of readytodrink coffee brews. Journal of Agricultural and Food Chemistry 51 (4): 942946.

Ankrah, N. A. & AppiahOpong, R. (1999). Toxicity of low levels of methylglyoxal: depletion of blood glutathione and adverse effect on glucose tolerance in mice. Toxicology Letters 109 (12): 6167.

Anon (2004). Questions about coffee and health. Indian coffee 68 : 3134.

Arendash, G. W., Mori, T., Cao, C. H., Mamcarz, M., Runfeldt, M., Dickson, A., Rezai Zadeh, K., Tan, J., Citron, B. A., Lin, X. Y., Echeverria, V. & Potter, H. (2009). Caffeine reverses cognitive impairment and decreases brain amyloidbeta Levels in aged Alzheimer's disease mice. Journal of Alzheimers Disease 17 (3): 661680.

Arnoldi, A., Corain, E. A., Scaglioni, L. & Ames, J. M. (1997). New colored compounds from the Maillard reaction between xylose and lysine. Journal of Agricultural and Food Chemistry 45 (3): 650655.

Arya, M. & Rao, L. J. M. (2007). An impression of coffee carbohydrates. Critical Reviews in Food Science and Nutrition 47 (1): 5167.

Babior, B. M. (1997). Superoxide: A twoedged sword. Brazilian Journal of Medical and Biological Research 30 (2): 141155.

Bailey, R. G., Ames, J. M. & Monti, S. M. (1996). An analysis of nonvolatile reaction products of aqueous Maillard model systems at pH5, using reversedphase HPLC with diodearray detection. Journal of the Science of Food and Agriculture 72 : 97103.

162 Baker, S. S. & Baker, R. D. (1992). Antioxidant enzymes in the differentiated Caco2 cell line. In Vitro Cellular & Developmental BiologyAnimal 28A (910): 643647.

Baker, S. S. & Baker, R. D. (1993). Caco2 cell metabolism of oxygen derived radicals. Digestive Diseases and Sciences 38 (12): 22732280.

Barone, J. J. & Roberts, H. R. (1996). Caffeine consumption. Food and Chemical Toxicology 34 (1): 119129.

Barranco Quintana, J. L., Allam, M. F., Serrano Del Castillo, A. & FernandezCrehuet Navajas, R. (2007). Alzheimer's disease and coffee: a quantitative review. Neurological Research 29 (1): 9195.

Baselt, R. C. (2002). Caffeine. Disposition of toxic drugs and chemicals in man . Foster City, Biomedical Publications.

Bekedam, E. K., De Laat, M., Schols, H. A., Van Boekel, M. & Smit, G. (2007). Arabinogalactan proteins are incorporated in negatively charged coffee brew melanoidins. Journal of Agricultural and Food Chemistry 55 (3): 761768.

Bekedam, E. K., Loots, M. J., Schols, H. A., Van Boekel, M. & Smit, G. (2008a). Roasting effects on formation mechanisms of coffee brew melanoidins. Journal of Agricultural and Food Chemistry 56 (16): 71387145.

Bekedam, E. K., Roos, E., Schols, H. A., van Boekel, M. & Smit, G. (2008b). Low molecular weight melanoidins in coffee brew. Journal of Agricultural and Food Chemistry 56 (11): 40604067.

Bekedam, E. K., Schols, H. A., Caemmerer, B., Kroh, L. W., van Boekel, M. & Smit, G. (2008c). Electron spin resonance (ESR) studies on the formation of roasting induced antioxidative structures in coffee brews at different degrees of roast. Journal of Agricultural and Food Chemistry 56 (12): 45974604.

Bekedam, E. K., Schols, H. A., Van Boekel, M. & Smit, G. (2006). High molecular weight melanoidins from coffee brew. Journal of Agricultural and Food Chemistry 54 (20): 76587666.

Bekedam, E. K., Schols, H. A., Van Boekel, M. & Smit, G. (2008d). Incorporation of chlorogenic acids in coffee brew melanoidins. Journal of Agricultural and Food Chemistry 56 (6): 20552063.

Belitz, H. D. & Grosch, W. (1999). Coffee, tea, cocoa. Food Chemistry . Berlin, Springer Verlag : 874883.

Boekel, M. A. J. S. v. (2001). Kinetic aspects of the Maillard reaction: a critical review. Nahrung/Food 45 (3): 150159.

163 Borrelli, R. C. & Fogliano, V. (2005). Bread crust melanoidins as potential prebiotic ingredients. Molecular Nutrition and Food Research 49 (7): 673678.

Borrelli, R. C., Visconti, A., Mennella, C., Anese, M. & Fogliano, V. (2002). Chemical characterization and antioxidant properties of coffee melanoidins. Journal of Agricultural and Food Chemistry 50 (22): 65276533.

Bossis, G. & Melchior, F. (2006). Regulation of SUMOylation by reversible oxidation of SUMO conjugating enzymes. Molecular Cell 21 (3): 349357.

Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein dye binding. Analytical Biochemistry 72 (12): 248254.

Budryn, G., Nebesny, E., Podsedek, A., Zyzelewicz, D., Materska, M., Jankowski, S. & Janda, B. (2009). Effect of different extraction methods on the recovery of chlorogenic acids, caffeine and Maillard reaction products in coffee beans. European Food Research and Technology 228 (6): 913922.

Cadden, I. S. H., Partovi, N. & Yoshida, E. M. (2007). Review article: Possible beneficial effects of coffee on liver disease and function. Alimentary Pharmacology & Therapeutics 26 (1): 17.

Caemmerer, B. & Kroh, L. W. (2006). Antioxidant activity of coffee brews. European Food Research and Technology 223 (4): 469474.

Cämmerer, B. & Kroh, L. W. (1995). Investigation of the influence of reaction conditions on the elementary composition of melanoidins. Food Chemistry 53 (1): 5559.

Cao, G. H., Alessio, H. M. & Cutler, R. G. (1993). Oxygen radical absorbance capacity assay for antioxidants Free Radical Biology and Medicine 14 (3): 303311.

Cao, G. H., Sofic, E. & Prior, R. L. (1996). Antioxidant capacity of tea and common vegetables. Journal of Agricultural and Food Chemistry 44 (11): 34263431.

Carrillo, J. A. & Benitez, J. (2000). Clinically significant pharmacokinetic interactions between dietary caffeine and medications. Clinical Pharmacokinetics . 39 : 127 153.

Castelluccio, C., Paganga, G., Melikian, N., Bolwell, G. P., Pridham, J., Sampson, J. & Riceevans, C. (1995). Antioxidant potential of intermediates in phenylpropanoid metabolism in higher plants. Febs Letters 368 (1): 188192.

Cavin, C., Holzhaeuser, D., Scharf, G., Constable, A., Huber, W. W. & Schilter, B. (2002). Cafestol and kahweol, two coffee specific diterpenes with anticarcinogenic activity. Food and Chemical Toxicology 40 (8): 11551163.

164 Cavin, C., MarinKuan, M., Langouet, S., Bezencon, C., Guignard, G., Verguet, C., Piguet, D., Holzhauser, D., Cornaz, R. & Schilter, B. (2008). Induction of Nrf2 mediated cellular defenses and alteration of phase I activities as mechanisms of chemoprotective effects of coffee in the liver. Food and Chemical Toxicology 46 (4): 12391248.

Cepinskas, G., Kvietys, P. R. & Aw, T. Y. (1994). Omega (3)lipid peroxides injure Caco2 cells relationship to the development of reduced glutathione antioxidant systems. Gastroenterology 107 (1): 8086.

Charrier, A. (1975). Variation of the caffeine content in coffee plants. International Colloquium Chimica Cafes 7: 295.

Chen, X.M. & Kitts, D. D. (2008a). Determining conditions for nitric oxide synthesis in Caco2 cells using Taguchi and factorial experimental designs. Analytical Biochemistry 381 (2): 185192.

Chen, X. M. & Kitts, D. D. (2008b). Antioxidant activity and chemical properties of crude and fractionated Maillard reaction products derived from four sugaramino acid Maillard reaction model systems. Maillard Reaction: Recent Advances in Food and Biomedical Sciences 1126 : 220224.

Chiu, W. K., Tanaka, M., Nagashima, Y. & Taguchi, T. (1991). Prevention of sardine lipid oxidation by antioxidative Maillard reaction products prepared from fructose tryptophan. Nippon Suisan Gakkaishi 57 (9): 17731781.

Choudhary, D., Chandra, D. & Kale, R. K. (1997). Influence of methylglyoxal on antioxidant enzymes and oxidative damage. Toxicology Letters 93 (23): 141152.

Chow, C. K. (1988). Interrelationships of cellular antioxidant defense systems Cellular antioxidant defense mechanisms . C. K. Chow. Boca Raton, FL, CRC Press. II.

Chowdhury, I., Mo, Y. Q., Gao, L., Kazi, A., Fisher, A. B. & Feinstein, S. I. (2009). Oxidant stress stimulates expression of the human peroxiredoxin 6 gene by a transcriptional mechanism involving an antioxidant response element. Free Radical Biology and Medicine 46 (2): 146153.

Chu, S. C., MarksKonczalik, J., Wu, H. P., Banks, T. C. & Moss, J. (1998). Analysis of the cytokinestimulated human inducible nitric oxide synthase (iNOS) gene: Characterization of differences between human and mouse iNOS promoters. Biochemical and Biophysical Research Communications 248 (3): 871878.

Claiborne, A. (1985). Catalase activity. Handbook of methods for oxygen radical research. R. Greenwald. Boca Raton, Fla, CRC Press : 283284.

Clifford, M. N. (1985). Coffee. Chemistry . R. J. Clarke and R. Macrae. London, UK, Elaevier Applied Science Publications. Vol. 1.

165 Clifford, M. N. (1999). Chlorogenic acids and other cinnamates nature, occurrence and dietary burden. Journal of the Science of Food and Agriculture 79 (3): 362372.

Corrao, G., Zambon, A., Bagnardi, V., D'Amicis, A., Klatsky, A. & Collaborative, S. G. (2001). Coffee, caffeine, and the risk of liver cirrhosis. Annals of Epidemiology 11 (7): 458465.

Crawford, D. (2002). Regulation of mammalian gene expression by reactive oxygen species. Reactive Oxygen Species in Biological Systems : 155171.

Crawford, D., Zbinden, I., Amstad, P. & Cerutti, P. (1988). Oxidant stress induces the protooncogenes cisfos and cismyc in mouse epidermal cells Oncogene 3(1): 27 32.

Cremin, P., KasimKarakas, S. & Waterhouse, A. L. (2001). LC/ESMS detection of hydroxycinnamates in human plasma and urine. Journal of Agricultural and Food Chemistry 49 (4): 17471750.

Cunha, R. A. (2008). Caffeine, adenosine receptors, memory and Alzheimer disease. Medicina Clinica 131 (20): 790795.

Daglia, M., Papetti, A., Gregotti, C., Berte, F. & Gazzani, G. (2000). In vitro antioxidant and ex vivo protective activities of green and roasted coffee. Journal of Agricultural and Food Chemistry 48 (5): 14491454.

Daglia, M., Racchi, M., Papetti, A., Lanni, C., Govoni, S. & Gazzani, G. (2004). In vitro and ex vivo antihydroxyl radical activity of green and roasted coffee. Journal of Agricultural and Food Chemistry 52 (6): 17001704.

De Roos, B., Meyboom, S., KosmeijerSchuil, T. G. & Katan, M. B. (1998). Absorption and urinary excretion of the coffee diterpenes cafestol and kahweol in healthy ileostomy volunteers. Journal of Internal Medicine 244 (6): 451460.

Deguine, V., Menasche, M., Ferrari, P., Fraisse, L., Pouliquen, Y. & Robert, L. (1998). Free radical depolymerization of hyaluronan by maillard reaction products: Role in liquefaction of aging vitreous. International Journal of Biological Macromolecules 22 (1): 1722.

del Castillo, M. D., Ames, J. M. & Gordon, M. H. (2002). Effect of roasting on the antioxidant activity of coffee brews. Journal of Agricultural and Food Chemistry 50 (13): 36983703.

del Castillo, M. D., Gordon, M. H. & Ames, J. M. (2005). Peroxyl radicalscavenging activity of coffee brews. European Food Research and Technology 221 (34): 471477.

166 DelgadoAndrade, C. & Morales, F. J. (2005). Unraveling the contribution of melanoidins to the antioxidant activity of coffee brews. Journal of Agricultural and Food Chemistry 53 (5): 14031407.

DelgadoAndrade, C., RufianHenares, J. A. & Morales, F. J. (2005). Assessing the antioxidant activity of melanoidins from coffee brews by different antioxidant methods. Journal of Agricultural and Food Chemistry 53 (20): 78327836.

DeMaria, C. A. B., Trugo, L. C., Neto, F. R. A., Moreira, R. F. A. & Alviano, C. S. (1996). Composition of green coffee watersoluble fractions and identification of volatiles formed during roasting. Food Chemistry 55 (3): 203207.

Devary, Y., Gottlieb, R. A., Lau, L. F. & Karin, M. (1991). Rapid and preferential activation of the cjun gene during the mammalian UV response Molecular and Cellular Biology 11 (5): 28042811.

Devasagayam, T. P. A., Kamat, J. P., Mohan, H. & Kesavan, P. C. (1996). Caffeine as an antioxidant: inhibition of lipid peroxidation induced by reactive oxygen species. Biochimica et Biophysica Acta (BBA) Biomembranes 1282 (1): 6370.

DinkovaKostova, A. T., Holtzclaw, W. D., Cole, R. N., Itoh, K., Wakabayashi, N., Katoh, Y., Yamamoto, M. & Talalay, P. (2002). Direct evidence that sulfhydryl groups of Keap1 are the sensors regulating induction of phase 2 enzymes that protect against carcinogens and oxidants. Proceedings of the National Academy of Sciences of the United States of America 99 (18): 1190811913.

Dunwiddie, T. V. & Masino, S. A. (2001). The role and regulation of adenosine in the central nervous system. Annual Review of Neuroscience 24 : 3155.

Eiserich, J. P. & Shibamoto, T. (1994). Antioxidant activity of volatile heterocyclic compounds. Journal of Agricultural and Food Chemistry 42 (5): 10601063.

Ekmekcioglu, C., StraussBlasche, G., Leibetseder, V. J. & Marktl, W. (1999). Toxicological and biochemical effects of different beverages on human intestinal cells. Food Research International 32 (6): 421427.

Elgawish, A., Glomb, M., Friedlander, M. & Monnier, V. M. (1996). Involvement of hydrogen peroxide in collagen crosslinking by high glucose in vitro and in vivo. Journal of Biological Chemistry 271 (22): 1296412971.

Ellenhorn, M. J. & Barceloux, D. (1988). Medical toxicology, diagnosis and treatment of human poisoning. New York, Elsevier.

Erbersdobler, H. F. & Faist, V. (2001). Metabolic transit of Amadori products. Nahrung Food 45 (3): 177181.

Esposito, F., Morisco, F., Verde, V., Ritieni, A., Alezio, A., Caporaso, N. & Fogliano, V. (2003). Moderate coffee consumption increases plasma glutathione but not

167 homocysteine in healthy subjects. Alimentary Pharmacology and Therapeutics 17 (4): 595601.

Faist, V. & Erbersdobler, H. F. (2001). Metabolic transit and in vivo effects of melanoidins and precursor compounds deriving from the Maillard reaction. Annals of Nutrition and Metabolism 45 : 112.

Faist, V. & Erbersdobler, H. F. (2002). Health impact of foodderived Maillard reaction products. Kieler Milchwirtschaftliche Forschungsberichte 54 (2): 137147.

Faist, V., Krome, K., Ames, J., Erbersdobler, H. F. (2001). Effects of nonenzymatic browning products formed by roasting of glucose/glycine and glucose/casein mixtures on NADPH cytochrome creductase and glutathioneStransferase in Caco2 Cells. Melanoidins in Food and Health . J. Ames, EUR19881. 2.

Faist, V., Lindenmeier, M., Geisler, C., Erbersdobler, H. F. & Hofmann, T. (2001). Influence of Molecular Weight Fractions Isolated from Roasted Malt on the Enzyme Activities of NADPH−Cytochrome c−Reductase and GlutathioneS transferase in Caco2 Cells. Journal of Agricultural and Food Chemistry 50 (3): 602606.

Farah, A. & Donangelo, C. M. (2006). Phenolic compounds in coffee. Brazilian Journal of Plant Physiology 18 (1): 2336.

Favreau, L. V. & Pickett, C. B. (1991). Transcriptional regulation of the rat NAD(P)H:quinone reductase gene. Identification of regulatory elements controlling basal level expression and inducible expression by planar aromatic compounds and phenolic antioxidants. Journal of Biological Chemistry 266 (7): 45564561.

Feng, R. T., Lu, Y. J., Bowman, L. L., Qian, Y., Castranova, V. & Ding, M. (2005). Inhibition of activator protein1, NFkappa B, and MAPKs and induction of phase 2 detoxifying enzyme activity by chlorogenic acid. Journal of Biological Chemistry 280 (30): 2788827895.

Fisher, A. B., Dodia, C., Manevich, Y., Chen, J. W. & Feinstein, S. I. (1999). Phospholipid hydroperoxidase are substrates for nonselenium glutathione peroxidase. Journal of Biological Chemistry 274 (30): 2132621334.

Franco, A. A., Odom, R. S. & Rando, T. A. (1999). Regulation of antioxidant enzyme gene expression in response to oxidative stress and during differentiation of mouse skeletal muscle. Free Radical Biology and Medicine 27 (910): 11221132.

Friedman, M. (2005). Biological effects of Maillard browning products that may affect acrylamide safety in food.

168 Fuster, M. D., Mitchell, A. E., Ochi, H. & Shibamoto, T. (2000). Antioxidative activities of heterocyclic compounds formed in brewed coffee. Journal of Agricultural and Food Chemistry 48 (11): 56005603.

George, S. E., Ramalakshmi, K. & Rao, L. J. M. (2008). A perception on health benefits of coffee. Critical Reviews in Food Science and Nutrition 48 (5): 464486.

Giovannucci, E. (1998). Metaanalysis of coffee consumption and risk of colorectal cancer. American Journal of Epidemiology 147 (11): 10431052.

Gokmen, V. & Senyuva, H. Z. (2006). Study of color and acrylamide formation in coffee, wheat flour and potato chips during heating. Food Chemistry 99 (2): 238243.

GomezRuiz, J. A., Ames, J. M. & Leake, D. S. (2008). Antioxidant activity and protective effects of green and dark coffee components against human low density lipoprotein oxidation. European Food Research and Technology 227 (4): 1017 1024.

Gonthier, M.P., Verny, M.A., Besson, C., Remesy, C. & Scalbert, A. (2003). Chlorogenic acid bioavailability largely depends on its metabolism by the gut microflora in rats. Journal of Nutrition 133 (6): 18531859.

Goya, L., DelgadoAndrade, C., RufianHenares, J. A., Bravo, L. & Morales, F. J. (2007). Effect of coffee Melanoidin on human hepatoma HepG2 cells. Protection against oxidative stress induced by tertbutylhydroperoxide. Molecular Nutrition & Food Research 51 (5): 536545.

Greer, F., Hudson, R., Ross, R. & Graham, T. (2001). Caffeine ingestion decreases glucose disposal during a hyperinsulinemiceuglycemic clamp in sedentary humans. Diabetes 50 (10): 23492354.

Gross, G., Jaccaud, E. & Huggett, A. C. (1997). Analysis of the content of the diterpenes cafestol and kahweol in coffee brews. Food and Chemical Toxicology 35 (6): 547 554.

Gu, F., Kim, J. M., Hayat, K., Xia, S., Feng, B. & Zhang, X. (2009). Characteristics and antioxidant activity of ultrafiltrated Maillard reaction products from a casein glucose model system. Food Chemistry 117 (1): 4854.

Guillot, F. L., Malnoe, A. & Stadler, R. H. (1996). Antioxidant properties of novel tetraoxygenated phenylindan isomers formed during thermal decomposition of caffeic acid. Journal of Agricultural and Food Chemistry 44 (9): 25032510.

Hancock, J. T., Desikan, R. & Neill, S. J. (2001). Role of reactive oxygen species in cell signalling pathways. Biochemical society transactions 29 : 345350.

Harland, B. F. (2000). Caffeine and nutrition. Nutrition 16 (78): 522526.

169 Harris, E. D. (1992). Copper as a cofactor and regulator of copper, zinc superoxide dismutase. Journal of Nutrition 122 (3): 636640.

Hartsfield, C. L., Alam, J., Cook, J. L. & Choi, A. M. K. (1997). Regulation of heme oxygenase1 gene expression in vascular smooth muscle cells by nitric oxide. American Journal of PhysiologyLung Cellular and Molecular Physiology 273 (5): L980L988.

Hayase, F., Takahashi, Y., Tominaga, S., Miura, M., Gomyo, T. & Kato, H. (1999). Identification of blue pigment formed in a Dxyloseglycine reaction system. Bioscience Biotechnology and Biochemistry 63 (8): 15121514.

Hayashi, T. & Namiki, M. (1986). Role of Sugar Fragmentation in an Early Stage Browning of Aminocarbonyl reaction of sugar with amino acid. Agricultural and Biological Chemistry 50 (8): 19651970.

Hayes, J. D., Ellis, E. M., Neal, G. E., Harrison, D. J. & Manson, M. M. (1999). Cellular response to cancer chemopreventive agents: contribution of the antioxidant responsive element to the adaptive response to oxidative and chemical stress. Biochemical Society Symposium 64 : 14168.

Hegele, J., Munch, G. & Pischetsrieder, M. (2009). Identification of hydrogen peroxide as a major cytotoxic component in Maillard reaction mixtures and coffee. Molecular Nutrition and Food Research 53 (6): 760769.

Higgins, L. G., Cavin, C., Toh, K., Yamamoto, M. & Hayes, J. D. (2008). Induction of cancer chemopreventive enzymes by coffee is mediated by transcription factor Nrf2. Evidence that the coffeespecific diterpenes cafestol and kahweol confer protection against acrolein. Toxicology and Applied Pharmacology 226 (3): 328 337.

Hodge, J. E. (1953). Dehydrated foods chemistry of browning reactions in model systems. . Journal of Agricultural and Food Chemistry 1(15): 928943.

Hofmann, T. (1997). Determination of the chemical structure of novel colored 1Hpyrrol 3(2H)one derivatives formed by Maillardtype reactions. Helvetica Chimica Acta 80 (6): 18431856.

Hofmann, T. (1998a). Characterization of the most intense coloured compounds from Maillard reactions of pentoses by application of color dilution analysis. Carbohydrate Research 313 (34): 203213.

Hofmann, T. (1998b). Studies on the relationship between molecular weight and the color potency of fractions obtained by thermal treatment of glucose amino acid and glucose/protein solutions by using ultracentrifugation and color dilution techniques. Journal of Agricultural and Food Chemistry 46 (10): 38913895.

170 Hofmann, T., Ames, J., Krome, K. & Faist, V. (2001). Determination of the molecular weight distribution of nonenzymatic browning products formed by roasting of glucose and glycine and studies on their effects on NADPHcytochrome c reductase and glutathioneStransferase in Caco2 cells. NahrungFood 45 (3): 189194.

Holland, R., Navamal, M., Velayutham, M., Zweier, J. L., Kensler, T. W. & Fishbein, J. C. (2009). Hydrogen peroxide Is a second messenger in phase 2 enzyme induction by cancer chemopreventive dithiolethiones. Chemical Research in Toxicology 22 (8): 14271434.

Ibarz, A., Garvín, A., Garza, S. & Pagán, J. (2009). Toxic effect of melanoidins from glucoseasparagine on trypsin activity. Food and Chemical Toxicology 47 (8): 20712075.

Jaiswal, A. K. (2004). Nrf2 signaling in coordinated activation of antioxidant gene expression. Free Radical Biology and Medicine 36 (10): 11991207.

Jing, H. & Kitts, D. D. (2002). Chemical and biochemical properties of caseinsugar Maillard reaction products. Food and Chemical Toxicology 40 (7): 10071015.

Jing, H. & Kitts, D. D. (2003). Chemistry of Maillard reaction products. Recent Res. Devel. Mol. Cell. Biochem 1: 7795.

Jing, H. & Kitts, D. D. (2004a). Antioxidant activity of sugarlysine Maillard reaction products in cell free and cell culture systems. Archives of Biochemistry and Biophysics 429 (2): 154163.

Jing, H. & Kitts, D. D. (2004b). Redoxrelated cytotoxic responses to different casein glycation products in Caco2 and Int407 cells. Journal of Agricultural and Food Chemistry 52 (11): 35773582.

Karakaya, S., El, S. N. & Tas, A. A. (2001). Antioxidant activity of some foods containing phenolic compounds. International Journal of Food Sciences and Nutrition 52 (6): 501508.

Kawakishi, S., Cheng, R. Z., Sato, S. & Uchida, K. (1994). Maillard Reactions in Chemistry, Food and Health. T. P. Labuza, G. A. Reineccius, V. M. Monnier, J. O'Brien and J. W. Baynes, The Royal Society of Chemistry, Cambridge.

Kawane, M., Yoshiki, Y. & Tsunakakawa, M. (1999). Radical scavenging activity of Maillard reaction substances. Phytochemicals and phytopharmaceuticals . F. Shahidi and C. T. Ho. Champaign, AOCS Press : 252260.

Kitts, D. D. (1997). An evaluation of the multiple effects of the antioxidant vitamins. Trends in Food Science and Technology 8(6): 198203.

171 Kitts, D. D. & Hu, C. (2005). Biological and chemical assessment of antioxidant activity of sugarlysine model Maillard reaction products. Maillard Reaction: Chemistry at the Interface of Nutrition, Aging, and Disease . J. W. Baynes, V. M. Monnier, J. M. Ames and S. R. Thorpe. New York, New York Acad Sciences. 1043 : 501512.

Kitts, D. D. & Wijewickreme, A. N. (1994). Effect of dietary caffeic and chlorogenic acids on in vivo xenobiotic enzyme systems. Plant Foods for Human Nutrition (Formerly Qualitas Plantarum) 45 (3): 287298.

Kitts, D. D., Wu, C. H. & Powrie, W. D. (1993). Effect of glucose lysine Maillard reaction product fractions on tissue xenobiotic enzyme systems. Journal of Agricultural and Food Chemistry 41 (12): 23592363.

Kurosaki, Y., Sato, H. & Mizugaki, M. (1989). Extraweak chemiluminescence of drugs. VIII. Extraweak chemiluminescence arising from the amadoricarbonyl reaction. Journal of Bioluminescence and Chemiluminescence 3: 1319.

Lam, L. K. T., Sparnins, V. L. & Wattenberg, L. W. (1987). Effects of derivatives of kahweol and cafestol on the activity of glutathione Stransferase in mice. Journal of Medicinal Chemistry 30 (8): 13991403.

Larsson, S. C. & Wolk, A. (2007). Coffee consumption and risk of liver cancer: A meta analysis. Gastroenterology 132 (5): 17401745.

Leanderson, P. & Tagesson, C. (1990). Cigarette smokeinduced DNAdamage: Role of hydroquinone and catechol in the formation of the oxidative DNAadduct, 8 hydroxydeoxyguanosine. ChemicoBiological Interactions 75 (1): 7181.

Leclere, J. & BirlouezAragon, I. (2001). The fluorescence of advanced Maillard products is a good indicator of lysine damage during the Maillard reaction. Journal of Agricultural and Food Chemistry 49 (10): 46824687.

Lee, C. (2000). Antioxidant ability of caffeine and its metabolites based on the study of oxygen radical absorbing capacity and inhibition of LDL peroxidation. Clinica Chimica Acta 295 (12): 141154.

Lee, J. M., Calkins, M. J., Chan, K. M., Kan, Y. W. & Johnson, J. A. (2003). Identification of the NFE2related factor2dependent genes conferring protection against oxidative stress in primary cortical astrocytes using oligonucleotide microarray analysis. Journal of Biological Chemistry 278 (14): 1202912038.

Lee, J. M., Moehlenkamp, J. D., Hanson, J. M. & Johnson, J. A. (2001). Nrf2dependent activation of the antioxidant responsive element by tertbutylhydroquinone is independent of oxidative stress in IMR32 human neuroblastoma cells. Biochemical and Biophysical Research Communications 280 (1): 286292.

172 Leong, L. P. & Wedzicha, B. L. (2000). A critical appraisal of the kinetic model for the Maillard browning of glucose with glycine. Food Chemistry 68 (1): 2128.

Lercker, G., Frega, N., Bocci, F. & RodriguezEstrada, M. (1995). High resolution gas chromatographic determination of diterpenic alcohols and sterols in coffee lipids. Chromatographia 41 (1): 2933.

Lerici, C. R., Barbanti, D., Manzano, M. & Cherubin, S. (1990). Early indicators of chemical changes in foods due to enzymatic or non enzymatic browning reactions. Study on heat treated model systems. LebensmittelWissenschaft & Technologie 23 (4): 289294.

Levin, R. E. (1982). Influence of caffeine on mutations induced by nitrosoguanidine in salmonella typhimurium tester strains. Environmental Mutagenesis 4(6): 689694.

Li, Y. & Jaiswal, A. K. (1994). Human antioxidantresponseelementmediated regulation of type 1 NAD(P)H:quinone oxidoreductase gene expression. Effect of sulfhydryl modifying agents. European Journal of Biochemistry 226 (1): 3139.

Lindenmeier, M., Faist, V. & Hofmann, T. (2002). Structural and functional characterization of pronyllysine, a novel protein modification in bread crust melanoidins showing in vitro antioxidative and phase I/II enzyme modulating activity. Journal of Agricultural and Food Chemistry 50 (24): 69977006.

Lingnert, H. (1990). Development of the Maillard reaction during food processing. Maillard Reaction in Food Processing, Human Nutrition and Physiology . P. A. Finot, H. U. Aeschbacher, R. F. Hurrell and R. Liardon : 171185.

Liu, S. F., Ye, X. B. & Malik, A. B. (1997). In vivo inhibition of nuclear factorkappa B activation prevents inducible nitric oxide synthase expression and systemic hypotension in a rat model of septic shock. Journal of Immunology 159 (8): 3976 3983.

Lowenstein, C. J., Glatt, C. S., Bredt, D. S. & Snyder, S. H. (1992). Cloned and expressed macrophage nitric oxide synthase contrasts with the brain enzyme Proceedings of the National Academy of Sciences of the United States of America 89 (15): 67116715.

MacDougall, D. B. & Granov, M. (1998). Relationship between Ultraviolet and Visible Spectra in Maillard Reactions and CIELAB colour space and visual appearance. The Maillard reaction in foods and medicine . J. O'Brien, H. E. Nursten, M. J. C. Crabbe and J. M. Ames. UK, Cambridge : 160165.

Macku, C. & Shibamoto, T. (1991). Volatile Antioxdiants produced from heated corn oil glycine model system. Journal of Agricultural and Food Chemistry 39 (11): 1990 1993.

173 Macrea, R. (1985). Nitrogenous components. Coffee Volume 1: Chemistry . R. J. Clarke and R. Macrea. London, Elsevier Applied Science Publishers Ltd. : 115152.

Maillard, L. C. (1912). The action of amino acids on sugar; The formation of melanoidin by a methodic route. Comptes Rendus Hebdomadaires Des Seances De L Academie Des Sciences 154 : 6668.

Manevich, Y. & Fisher, A. B. (2005). Peroxiredoxin 6, a 1Cys peroxiredoxin, functions in antioxidant defense and lung phospholipid metabolism. Free Radical Biology and Medicine 38 (11): 14221432.

Manna, C., Galletti, P., Cucciolla, V., Moltedo, O., Leone, A. & Zappia, V. (1997). The protective effect of the olive oil polyphenol (3,4dihydroxyphenyl)ethanol counteracts reactive oxygen metaboliteinduced cytotoxicity in Caco2 cells. Journal of Nutrition 127 (2): 286292.

Masaki, H., Okano, Y. & Sakurai, H. (1998). Differential role of catalase and glutathione peroxidase in cultured human fibroblasts under exposure of H2O2 or ultraviolet B light. Archives of Dermatological Research 290 (3): 113118.

Massini, R., Nicoli, M. C., Cassara, A. & Lerici, C. R. (1990). Study on physical and physicochemical changes of coffee beans during roasting. Italian Journal of Food Science 2(2): 123130.

Mastrocola, D. & Munari, M. (2000). Progress of the Maillard reaction and antioxidant action of Maillard reaction products in preheated model systems during storage. Journal of Agricultural and Food Chemistry 48 (8): 35553559.

Mateos, R., Goya, L. & Bravo, L. (2006). Uptake and metabolism of hydroxycinnamic acids (chlorogenic, caffeic, and ferulic acids) by HepG2 cells as a model of the human liver. Journal of Agricultural and Food Chemistry 54 (23): 87248732.

Mayer, J. M. (2004). Protoncoupled electron transfer: A reaction chemist's view. Annual Review of Physical Chemistry 55 : 363390.

McCusker, R. R., Goldberger, B. A. & Cone, E. J. (2003). Technical note: Caffeine content of specialty coffees. Journal of Analytical Toxicology 27 : 520522.

Michalska, A., AmigoBenavent, M., Zielinski, H. & del Castillo, M. D. (2008). Effect of bread making on formation of Maillard reaction products contributing to the overall antioxidant activity of rye bread. Journal of Cereal Science 48 (1): 123 132.

Minotti, G. & Aust, S. D. (1992). Redox cycling of iron and lipidperoxidation Lipids 27 (3): 219226.

174 Mochizuki, M., Yamazaki, S., Kano, K. & Ikeda, T. (2002). Kinetic analysis and mechanistic aspects of autoxidation of catechins. Biochimica et Biophysica Acta General Subjects 1569 (13): 3544.

Monteiro, M., Farah, A., Perrone, D., Trugo, L. C. & Donangelo, C. (2007). Chlorogenic acid compounds from coffee are differentially absorbed and metabolized in humans. Journal of Nutrition 137 (10): 21962201.

Monti, S. M., Ritieni, A., Graziani, G., Randazzo, G., Mannina, L., Segre, A. L. & Fogliano, V. (1999). LC/MS analysis and antioxidative efficiency of Maillard reaction products from a lactoselysine model system. Journal of Agricultural and Food Chemistry 47 (4): 15061513.

Morales, F. J. (2005). Assessing the nonspecific hydroxyl radical scavenging properties of melanoidins in a Fentontype reaction system. Analytica Chimica Acta 534 (1): 171176.

Morales, F. J., FernandezFraguas, C. & JimenezPerez, S. (2005). Ironbinding ability of melanoidins from food and model systems. Food Chemistry 90 (4): 821827.

Morales, F. J. & JimenezPerez, S. (2001). Free radical scavenging capacity of Maillard reaction products as related to colour and fluorescence. Food Chemistry 72 (1): 119125.

Morales, F. J. & JimenezPerez, S. (2004). Peroxyl radical scavenging activity of melanoidins in aqueous systems. European Food Research and Technology 218 (6): 515520.

Morales, F. J. & van Boekel, M. (1998). A study on advanced Maillard reaction in heated casein/sugar solutions: Colour formation. International Dairy Journal 8(1011): 907915.

Mosmann, T. (1983). Rapid colormetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays Journal of Immunological Methods 65 (12): 5563.

Murakami, M., Shigeeda, A., Danjo, K., Yamaguchi, T., Takamura, H. & Matoba, T. (2002). Radicalscavenging activity and brightly colored pigments in the early stage of the Maillard reaction. Journal of Food Science 67 (1): 9396.

Murthy, U. M. N., Liang, Y. H., Kumar, P. P. & Sun, W. Q. (2002). Nonenzymatic protein modification by the Maillard reaction reduces the activities of scavenging enzymes in Vigna radiata. Physiologia Plantarum 115 (2): 213220.

Muscat, S., Pelka, J., Hegele, J., Weigle, B., Munch, G. & Pischetsrieder, M. (2007). Coffee and Maillard products activate NFkappa B in macrophages via H 2O2 production. Molecular Nutrition and Food Research 51 (5): 525535.

175 Nagao, M., Fujita, Y., Wakabayashi, K., Nukaya, H., Kosuge, T. & Sugimura, T. (1986). Mutagens in coffee and other beverages Environmental Health Perspectives 67 : 8991.

Nakanishi, I., Fukuhara, K., Ohkubo, K., Shimada, T., Kansui, H., Kurihara, M., Urano, S., Fukuzumi, S. & Miyata, N. (2001). Superoxide anion generation via electron transfer oxidation of catechin dianion by molecular oxygen in an aprotic medium. Chemistry Letters (11): 11521153.

Namiki, M. & Hayashi, T. (1975). Development of novel free radicals during amino carbonyl reaction of sugars with aminoacids Journal of Agricultural and Food Chemistry 23 (3): 487491.

Namiki, M. & Hayashi, T. (1983). A new mechanism of the Maillard reaction involving sugar fragmentation and free radical formation American Chemical Society Symposium Series 215 : 2146.

Nath, K. A., Grande, J., Croatt, A., Haugen, J., Kim, Y. & Rosenberg, M. E. (1998). Redox regulation of renal DNA synthesis, transforming growth factorbeta 1 and collagen gene expression. Kidney International 53 (2): 367381.

Nguyen, T., Rushmore, T. H. & Pickett, C. B. (1994). Transcriptional regulation of a rat liver glutathione Stransferase Ya subunit gene. Analysis of the antioxidant response element and its activation by the phorbol ester 12O tetradecanoylphorbol13acetate. Journal of Biological Chemistry 269 (18): 1365613662.

Nguyen, T., Sherratt, P. J. & Pickett, C. B. (2003). Regulatory mechanisms controlling gene expression mediated by the antioxidant response element. Annual Review of Pharmacology and Toxicology 43 : 233260.

Nicoli, M. C., Anese, M., Manzocco, L. & Lerici, C. R. (1997). Antioxidant properties of coffee brews in relation to the roasting degree. LebensmittelWissenschaft and Technologie 30 (3): 292297.

Nienaber, U. & Eichner, K. (1995). The antioxidative effect of Maillard reaction products in model systems and roasted hazelnuts. Fett Wissenschaft TechnologieFat Science Technology 97 (12): 435444.

Nose, K., Shibanuma, M., Kikuchi, K., Kageyama, H., Sakiyama, S. & Kuroki, T. (1991). Transcriptional activation of early response genes by hydrogen peroxide in a mouse osteoblastic cell line. European Journal of Biochemistry 201 (1): 99106.

Nunes, F. M. & Coimbra, M. A. (2007). Melanoidins from coffee infusions. Fractionation, chemical characterization, and effect of the degree of roast. Journal of Agricultural and Food Chemistry 55 (10): 39673977.

176 Nussler, A. K. & Billiar, T. R. (1993). Inflammation, immunoregulation, and inducible nitricoxide synthase. Journal of Leukocyte Biology 54 (2): 171178.

Obretenov, T., Kuntcheva, M. J. & Panchev, I. N. (1986). Influence of reaction conditions on forming nondialyzable melanoidins from glucose and glysine. Journal of Food Processing and Preservation 10 : 251268.

OBrien, J. & Morrissey, P. A. (1989). Nutritional and toxicological aspects of the Maillard browning reaction in foods. Critical Reviews in Food Science and Nutrition 28 (3): 211248.

Okano, J. I., Nagahara, T., Matsumoto, K. & Murawaki, Y. (2008). Caffeine inhibits the proliferation of liver cancer cells and activates the MEK/ERK/EGFR signalling pathway. Basic and Clinical Pharmacology and Toxicology 102 (6): 543551.

Oosterveld, A., Voragen, A. G. J. & Schols, H. A. (2003). Effect of roasting on the carbohydrate composition of Coffea arabica beans. Carbohydrate Polymers 54 (2): 183192.

Ou, B. X., HampschWoodill, M. & Prior, R. L. (2001). Development and validation of an improved oxygen radical absorbance capacity assay using fluorescein as the fluorescent probe. Journal of Agricultural and Food Chemistry 49 (10): 4619 4626.

Ou, B. X., Huang, D. J., HampschWoodill, M., Flanagan, J. A. & Deemer, E. K. (2002). Analysis of antioxidant activities of common vegetables employing oxygen radical absorbance capacity (ORAC) and ferric reducing antioxidant power (FRAP) assays: A comparative study. Journal of Agricultural and Food Chemistry 50 (11): 31223128.

Paglia, D. E. & Valentine, W. N. (1967). Studies on the quantitative and qualitative characterisation of erythtrocyte glutathione peroxidase. Journal of Laboratory and Clinical Medicine 70 (1): 158169.

Park, Y. S., Koh, Y. H., Takahashi, M., Miyamoto, Y., Suzuki, K., Dohmae, N., Takio, K., Honke, K. & Taniguchi, N. (2003). Identification of the binding site of methylglyoxal on glutathione peroxidase: Methylglyoxal inhibits glutathione peroxidase activity via binding to glutathione binding sites Arg 184 and 185. Free Radical Research 37 (2): 205211.

Pellegrini, N., Serafini, M., Colombi, B., Del Rio, D., Salvatore, S., Bianchi, M. & Brighenti, F. (2003). Total antioxidant capacity of plant foods, beverages and oils consumed in Italy assessed by three different in vitro assays. Journal of Nutrition 133 (9): 28122819.

Pichorner, H., Jessner, G. & Ebermann, R. (1993). tBOOH acts as a suicide substrate for catalase. Archives of Biochemistry and Biophysics 300 (1): 258264.

177 Pinkus, R., Weiner, L. M. & Daniel, V. (1996). Role of oxidants and antioxidants in the induction of AP1, NFkappa B, and glutathione Stransferase gene expression. Journal of Biological Chemistry 271 (23): 1342213429.

Pongor, S., Ulrich, P. C., Bencsath, F. A. & Cerami, A. (1984). Aging of protein: isolation and identification of a fluorescent chromophore from the reaction of polypeptides with glucose. Proceedings of the National Academy of Sciences of the United States of AmericaBiological Sciences 81 (9): 26842688.

Popovich, D. G. & Kitts, D. D. (2004a). Ginsenosides 20(S)protopanaxadiol and Rh2 reduce cell proliferation and increase subG, cells in two cultured intestinal cell lines, Int407 and Caco2. Canadian Journal of Physiology and Pharmacology 82 (3): 183190.

Popovich, D. G. & Kitts, D. D. (2004b). Mechanistic studies on protopanaxadiol, Rh2, and ginseng (Panax quinquefolius) extract induced cytotoxicity in intestinal Caco 2 cells. Journal of Biochemical and Molecular Toxicology 18 (3): 143149.

Powrie, W. D., Wu, C. H. & Molund, V. P. (1986). Browning reaction systems as sources of mutagens and antimutagens. Environmental Health Perspectives 67 : 4754.

Pulido, R., HernandezGarcia, M. & SauraCalixto, F. (2003). Contribution of beverages to the intake of lipophilic and hydrophilic antioxidants in the Spanish diet. European Journal of Clinical Nutrition 57 (10): 12751282.

PurdomDickinson, S. E., Sheveleva, E. V., Sun, H. P. & Chen, Q. M. (2007). Translational control of Nrf2 protein in activation of antioxidant response by oxidants. Molecular Pharmacology 72 : 10741081.

Putnam, C. D., Arvai, A. S., Bourne, Y. & Tainer, J. A. (2000). Active and inhibited human catalase structures: ligand and NADPH binding and catalytic mechanism. Journal of Molecular Biology 296 (1): 295309.

Radtke, J., Linseisen, J. & Wolfram, G. (1998). Phenolic acid intake of adults in a Bavarian subgroup of the national food consumption survey. Zeitschrift Fur Ernahrungswissenschaft 37 (2): 190197.

Raha, S. & Robinson, B. H. (2000). Mitochondria, oxygen free radicals, disease and ageing. Trends in Biochemical Sciences 25 (10): 502508.

Ranheim, T. & Halvorsen, B. (2005). Coffee consumption and human health beneficial or detrimental? Mechanisms for effects of coffee consumption on different risk factors for cardiovascular disease and Type 2 diabetes mellitus. Molecular Nutrition and Food Research 49 (3): 274284.

Ratnayake, W. M. N., Hollywood, R., O'Grady, E. & Stavric, B. (1993). Lipid content and composition of coffee brews prepared by different methods. Food and Chemical Toxicology 31 (4): 263269.

178 Rhee, S. G., Yang, K. S., Kang, S. W., Woo, H. A. & Chang, T. S. (2005). Controlled elimination of intracellular H2O2: Regulation of peroxiredoxin, catalase, and glutathione peroxidase via posttranslational modification. Antioxidants and Redox Signaling 7(56): 619626.

Richelle, M., Tavazzi, I. & Offord, E. (2001). Comparison of the antioxidant activity of commonly consumed polyphenolic beverages (coffee, cocoa, and tea) prepared per cup serving. Journal of Agricultural and Food Chemistry 49 (7): 34383442.

Rizzi, G. P. (1997). Chemical structure of colored Maillard reaction products. Food Reviews International 13 (1): 128.

Rizzi, G. P. (1999). Formation of sulfurcontaining volatiles under coffee roasting conditions. Abstracts of Papers American Chemical Society 217 (12): AGFD 48.

Roberts, R. L. & Lloyd, R. V. (1997). Free radical formation from secondary amines in the Maillard reaction. Journal of Agricultural and Food Chemistry 45 (7): 2413 2418.

RufianHenares, J. A. & Morales, F. J. (2007a). Antimicrobial activity of melanoidins. Journal of Food Quality 30 (2): 160168.

RufianHenares, J. A. & Morales, F. J. (2007b). Effect of in vitro enzymatic digestion on antioxidant activity of coffee melanoidins and fractions. Journal of Agricultural and Food Chemistry 55 : 1001610021.

RufianHenares, J. A. & Morales, F. J. (2007c). Functional properties of melanoidins: In vitro antioxidant, antimicrobial and antihypertensive activities. Food Research International 40 (8): 9951002.

Rushmore, T. H. & Pickett, C. B. (1993). Glutathione Stransferases, structure, regulation and therapeutic implications. Journal of Biological Chemistry 268 (16): 11475 11478.

Sablina, A. A., Budanov, A. V., Ilyinskaya, G. V., Agapova, L. S., Kravchenko, J. E. & Chumakov, P. M. (2005). The antioxidant function of the p53 tumor suppressor. Nature Medicine 11 (12): 13061313.

Sacchetti, G., Di Mattia, C., Pittia, P. & Mastrocola, D. (2009). Effect of roasting degree, equivalent thermal effect and coffee type on the radical scavenging activity of coffee brews and their phenolic fraction. Journal of Food Engineering 90 (1): 74 80.

Saenz, A. T., Elisia, I., Innis, S. M., Friel, J. K. & Kitts, D. D. (2009). Use of ORAC to assess antioxidant capacity of human milk. Journal of Food Composition and Analysis 22 (78): 694698.

179 Sakamoto, W., Isomura, H., Fujie, K., Nishihira, J., Ozaki, M. & Yukawa, S. (2003). Coffee increases levels of urinary 8hydroxydeoxyguanosine in rats. Toxicology 183 (13): 255263.

SanchezGonzalez, I., JimenezEscrig, A. & SauraCalixto, F. (2005). In vitro antioxidant activity of coffees brewed using different procedures (Italian, espresso and filter). Food Chemistry 90 (12): 133139.

Savolainen, H. (1992). Tannin content of tea and coffee Journal of Applied Toxicology 12 (3): 191192.

Scandalios, J. G. (2005). Oxidative stress: molecular perception and transduction of signals triggering antioxidant gene defenses. Brazilian Journal of Medical and Biological Research 38 (7): 9951014.

Seiquer, I., RuizRoca, B., Mesias, M., MunozHoyos, A., Galdo, G., Ochoa, J. J. & Navarro, M. P. (2008). The antioxidant effect of a diet rich in Maillard reaction products is attenuated after consumption by healthy male adolescents. In vitro and in vivo comparative study. Journal of the Science of Food and Agriculture 88 (7): 12451252.

Seo, H. S., Hirano, M., Shibato, J., Rakwal, R., Hwang, I. K. & Masuo, Y. (2008). Effects of coffee bean aroma on the rat brain stressed by sleep deprivation: A selected transcript and 2D gelbased proteome analysis. Journal of Agricultural and Food Chemistry 56 (12): 46654673.

Sies, H. (1993). Strategies of antioxidant defense. European Journal of Biochemistry 215 (2): 213219.

Sies, H. & Cadenas, E. (1985). Oxidative stressdamage to intact cells and organs. Philosophical Transactions of the Royal Society of London Series BBiological Sciences 311 (1152): 617631.

Singh, A., Rangasamy, T., Thimmulappa, R. K., Lee, H., Osburn, W. O., BrigeliusFlohe, R., Kensler, T. W., Yamamoto, M. & Biswal, S. (2006). Glutathione peroxidase 2, the major cigarette smokeinducible isoform of GPX in lungs, is regulated by Nrf2. American Journal of Respiratory Cell and Molecular Biology 35 (6): 639 650.

Somoza, V. (2005). Five years of research on health risks and benefits of Maillard reaction products: An update. Molecular Nutrition and Food Research 49 (7): 663672.

Somoza, V., Lindenmeier, M., Wenzel, E., Frank, O., Erbersdobler, H. F. & Hofmann, T. (2003). Activityguided identification of a chemopreventive compound in coffee beverage using in vitro and in vivo techniques. Journal of Agricultural and Food Chemistry 51 (23): 68616869.

180 Somoza, V., Wenzel, E., Lindenmeier, M., Grothe, D., Erbersdobler, H. F. & Hofmann, T. (2005). Influence of feeding malt,bread crust, and a pronylated protein on the activity of chemopreventive enzymes and antioxidative defense parameters in vivo. Journal of Agricultural and Food Chemistry 53 (21): 81768182.

Soriano, F. X., Leveille, F., Papadia, S., Higgins, L. G., Varley, J., Baxter, P., Hayes, J. D. & Hardingham, G. E. (2008). Induction of sulfiredoxin expression and reduction of peroxiredoxin hyperoxidation by the neuroprotective Nrf2 activator 3H1,2 dithiole3thione. Journal of Neurochemistry 107 (2): 533543.

Sun, Y. X., Hayakawa, S., Chuamanochan, M., Fujimoto, M., Innun, A. & Izumori, K. (2006). Antioxidant effects of Maillard reaction products obtained from ovalbumin and different Daldohexoses. Bioscience Biotechnology and Biochemistry 70 (3): 598605.

Svenningsson, P., Strom, A., Johansson, B. & Fredholm, B. B. (1995). Increased expression of cjun, junB, AP1, and preproenkephalin mRNA in rat striatum following a single injection of caffeine. Journal of Neuroscience 15 (5): 3583 3593.

Svilaas, A., Sakhi, A. K., Andersen, L. F., Svilaas, T., Strom, E. C., Jacobs, D. R., Ose, L. & Blomhoff, R. (2004). Intakes of antioxidants in coffee, wine, and vegetables are correlated with plasma carotenoids in humans. Journal of Nutrition 134 (3): 562 567.

Tak, P. P. & Firestein, G. S. (2001). NFkappa B: a key role in inflammatory diseases. Journal of Clinical Investigation 107 (1): 711.

Takenaka, M., Sato, N., Asakawa, H., Wen, X., Murata, M. & Homma, S. (2005). Characterization of a metalchelating substance in coffee. Bioscience Biotechnology and Biochemistry 69 (1): 2630.

Tanswell, A. K. & Freeman, B. A. (1984). Pulmonary antioxidant enzyme maturation in the fetal and neonatal rat. I. Developmental profiles. Pediatric Research 18 (9): 871874.

Tavani, A. & La Vecchia, C. (2004). Coffee, decaffeinated coffee, tea and cancer of the colon and rectum: a review of epidemiological studies, 19902003. Cancer Causes and Control 15 (8): 743757.

Thornalley, P. J., Wolff, S. P., Crabbe, M. J. C. & Stern, A. (1984). The oxidation of oxyhemoglobin by glyceraldehyde and other simple monosaccharides. . Biochemical Journal 217 (3): 615622.

Trugo, L. C. (1984). HPLC in coffee analysis. England, Univeristy of Reading. Doctorate thesis.

181 Tuohy, K. M., Probert, H. M., Smejkal, C. W. & Gibson, G. R. (2003). Using probiotics and prebiotics to improve gut health. Drug Discovery Today 8(15): 692700.

Ueda, Y., Miyata, T., Hashimoto, T., Yamada, H., Izuhara, Y., Sakai, H. & Kurokawa, K. (1998). Implication of altered redox regulation by antioxidant enzymes in the increased plasma pentosidine, an advanced glycation end product, in uremia. Biochemical and Biophysical Research Communications 245 (3): 785790.

Urgert, R., KosmeijerSchuil, T. G. & Katan, M. B. (1996). Intake levels, sites of action and excretion routes of the cholesterolelevating diterpenes from coffee beans in humans. Biochemical Society Transactions 24 (3): 800806.

Urgert, R., Vanderweg, G., Kosmeijerschuil, T. G., Vandebovenkamp, P., Hovenier, R. & Katan, M. B. (1995). Levels of the cholesterolelevating diterpenes cafestrol and kahweol in various coffee brews. Journal of Agricultural and Food Chemistry 43 (8): 21672172.

Vachon, P. H. & Beaulieu, J. F. (1992). Transient mosaic patterns of morphological and functional differentiation in the Caco2 cell line. Gastroenterology 103 (2): 414 423.

Valko, M., Leibfritz, D., Moncol, J., Cronin, M. T. D., Mazur, M. & Telser, J. (2007). Free radicals and antioxidants in normal physiological functions and human disease. The International Journal of Biochemistry and Cell Biology 39 (1): 4484. van Dam, R. M. & Hu, F. B. (2005). Coffee consumption and risk of Type 2 diabetes A systematic review. JamaJournal of the American Medical Association 294 (1): 97104.

Veal, E. A., Day, A. M. & Morgan, B. A. (2007). Hydrogen peroxide sensing and signaling. Molecular Cell 26 (1): 114.

Venugopal, R. & Jaiswal, A. K. (1998). Nrf2 and Nrf1 in association with Jun proteins regulate antioxidant response elementmediated expression and coordinated induction of genes encoding detoxifying enzymes. Oncogene 17 (24): 31453156.

Wagner, K. H., Derkits, S., Herr, M., Schuh, W. & Elmadfa, I. (2002). Antioxidative potential of melanoidins isolated from a roasted glucoseglycine model. Food Chemistry 78 (3): 375382.

Wang, Y. H., Crawford, D. R. & Davies, K. J. A. (1996). adapt33, a novel oxidant inducible RNA from hamster HA1 cells. Archives of Biochemistry and Biophysics 332 (2): 255260.

Waris, G. & Ahsan, H. (2006). Reactive oxygen species: role in the development of cancer and various chronic conditions. Journal of Carcinogenesis 5(1): 14.

182 Wijeratne, S. S. K., Cuppett, S. L. & Schlegel, V. (2005). Hydrogen peroxide induced oxidative stress damage and antioxidant enzyme response in Caco2 human colon cells. Journal of Agricultural and Food Chemistry 53 (22): 87688774.

Wijewickreme, A. N. & Kitts, D. D. (1997). Influence of reaction conditions on the oxidative behavior of model Maillardreaction products. Journal of Agricultural and Food Chemistry 45 (12): 45714576.

Wijewickreme, A. N. & Kitts, D. D. (1998a). Metal chelating and antioxidant activity of model Maillard reaction products. ProcessInduced Chemical Changes in Food 434 : 245254.

Wijewickreme, A. N. & Kitts, D. D. (1998b). Modulation of metalinduced cytotoxicity by Maillard reaction products isolated from coffee brew. Journal of Toxicology and Environmental HealthPart aCurrent Issues 55 (2): 151168.

Wijewickreme, A. N. & Kitts, D. D. (1998c). Modulation of metalinduced genotoxicity by Maillard reaction products isolated from coffee. Food and Chemical Toxicology 36 (7): 543553.

Wijewickreme, A. N., Kitts, D. D. & Durance, T. D. (1997). Reaction conditions influence the elementary composition and metal chelating affinity of nondialyzable model Maillard reaction products. Journal of Agricultural and Food Chemistry 45 (12): 45774583.

Wolff, S. P. & Dean, R. T. (1987). Glucose autoxidation and protein modificationThe potential role of autoxidative glycosylation in diabetes. Biochemical Journal 245 (1): 243250.

Wong, J. W. & Shibamoto, T. (1996). Genotoxicity of Maillard reaction products. The Maillard reaction: consequences for the chemical and life sciences . P. Ikan. England, John Wiley & Sons Ltd. : 129160.

Wyvratt, M. J. & Patchett, A. A. (1985). Recent developments in the design of angiotensinconverting enzyme inhibitors. Medicinal Research Reviews 5(4): 483 531.

Xie, Q. W., Kashiwabara, Y. & Nathan, C. (1994). Role of transcription factor NFkappa B/REL in induction of nitric oxide synthase. Journal of Biological Chemistry 269 (7): 47054708.

Yamanaka, N., Oda, O. & Nagao, S. (1997). Prooxidant activity of caffeic acid, dietary nonflavonoid phenolic acid, on Cu2+induced low density lipoprotein oxidation. FEBS Letters 405 (2): 186190.

Yanagimoto, K., Lee, K. G., Ochi, H. & Shibamoto, T. (2002). Antioxidative activity of heterocyclic compounds found in coffee volatiles produced by Maillard reaction. Journal of Agricultural and Food Chemistry 50 (19): 54805484.

183 Yang, S. P., Wilson, K., Kawa, A. & Raner, G. M. (2006). Effects of green tea extracts on gene expression in HepG2 and Cal27 cells. Food and Chemical Toxicology 44 (7): 10751081.

Yaylayan, V. A., Keyhani, A. & HuyguesDespointes, A. (1998). Generation and the fate of C2, C3, and C4 reactive fragments formed in Maillard model systems of [C 13]glucose and [C13]glycine or proline. ProcessInduced Chemical Changes in Food 434 : 237244.

Yen, G. C., Liao, C. M. & Wu, S. C. (2002). Influence of Maillard reaction products on DNA damage in human lymphocytes. Journal of Agricultural and Food Chemistry 50 (10): 29702976.

YetikAnacak, G. & Catravas, J. D. (2006). Nitric oxide and the endothelium: History and impact on cardiovascular disease. Vascular Pharmacology 45 (5): 268276.

Yuan, Y. V. & Kitts, D. D. (1997). Endogenous antioxidants: Role of antioxidant enzymes in biological system. Natural antioxidants: Chemistry, health effects, and applications : 258270.

Zhang, Y. S. & Gordon, G. B. (2004). A strategy for cancer prevention: Stimulation of the Nrf2ARE signaling pathway. Molecular Cancer Therapeutics 3(7): 885893.

Zhen, J. H., Lu, H., Wang, X. Q., Vaziri, N. D. & Zhou, X. J. (2008). Upregulation of endothelial and inducible nitric oxide synthase expression by reactive oxygen species. American Journal of Hypertension 21 (1): 2834.

Zheng, X. & Hendry, W., 3rd (1997). Neonatal diethylstilbestrol treatment alters the estrogenregulated expression of both cell proliferation and apoptosisrelated protooncogenes (cjun, cfos, cmyc, bax, bcl2, and bclx) in the hamster uterus. Cell Growth and Differentiation 8(4): 425434.

184 APPENDIX

Table 1 Effect of coffee extracts on the glutathione peroxidase activity in Caco2 cells 1 3 h 24 h 72 h Control 54+1 52+4 44+7 LR 56+10 46+3 42+5 DR 56+4 47+3 48+2 1 The values (mean + SD, n = 3) represent enzyme activity, expressed as mU/mg protein. Cells were exposed to 0.1 mg/ml light roasted (LR) and dark roasted (DR) coffee for 3 h, 24 h and 72 h. Cells incubated in culture medium were used as control.

Table 2 Effect of coffee extracts on the glutathione reductase activity in Caco2 cells 1 3 h 24 h 72 h Control 20+3 17+1 19+1 LR 21+2 19+2 21+2 DR 23+2 18+0 18+2 1 The values (mean + SD, n = 3) represent enzyme activity, expressed as mU/mg protein. Cells were exposed to 0.1 mg/ml light roasted (LR) and dark roasted (DR) coffee for 3 h, 24 h and 72 h. Cells incubated in culture medium were used as control.

Table 3 Effect of coffee extracts on the superoxide dismutase activity in Caco2 cells 1 3 h 24 h 72 h Control 5.1+0.2 5.0+0.3 4.8+0.2 LR 6.1+0.5 4.1+0.0 4.8+0.8 DR 6.0+0.6 4.3+0.6 5.1+0.5 1 The values (mean + SD, n = 3) represent enzyme activity, expressed as U/mg protein. Cells were exposed to 0.1 mg/ml light roasted (LR) and dark roasted (DR) coffee for 3 h, 24 h and 72 h. Cells incubated in culture medium were used as control.

Table 4 Effect of coffee extracts on the glutathione reductase activity in Caco2 cells 1 3 h 24 h 72 h Control 17+1 18+2 19+4 GB 16+2 20+2 19+4 LR 16+1 15+1 16+2 DR 16+1 18+2 18+5 1 The values (mean + SD, n = 3) represent enzyme activity, expressed as mU/mg protein. Cells were exposed to 1.0 mg/ml green coffee bean extract (GB), light roasted (LR) and dark roasted (DR) coffee for 3 h, 24 h and 72 h. Cells incubated in culture medium were used as control.

185

Table 5 Effect of coffee extracts on the superoxide dismutase activity in Caco2 cells 1 3 h 24 h 72 h Control 5.9+0.9 5.9+0.5 6.4+1.1 GB 5.7+0.3 5.7+0.2 6.1+0.4 LR 5.8+0.1 5.0+0.6 5.5+0.2 DR 5.6+0.4 5.0+0.3 5.7+0.4 1 The values (mean + SD, n = 3) represent enzyme activity, expressed as U/mg protein. Cells were exposed to 1.0 mg/ml green coffee bean extract (GB), light roasted (LR) and dark roasted (DR) coffee for 3 h, 24 h and 72 h. Cells incubated in culture medium were used as control.

Table 6 Effect of light roasted coffee extracts and fractions on Caco2 superoxide 1 dismutase activity with (+) or without () H 2O2 treatment

H 2O2 +H 2O2 Control 5.9+0.5 5.4+0.6 Defatted 5.0+0.6 5.6+0.3 Fraction II 5.0+0.4 4.8+0.3 Fraction III 5.3+0.6 6.2+0.6 1 The values (mean + SD, n = 3) represent enzyme activity, expressed as U/mg protein. Cells were exposed to 1.0 mg/ml light roasted coffee samples in culture medium for 24 h. Some cells were further treated with H 2O2 for 2 h. Cells incubated in culture medium were used as positive control. Cells treated with H2O2 added culture medium were used as positive control.

Table 7 Effect of green coffee bean extracts and fractions on Caco2 superoxide 1 dismutase activity with (+) or without () H 2O2 treatment

H 2O2 +H 2O2 Control 5.9+0.5 5.4+0.6 Defatted 5.7+0.2 5.3+0.2 Fraction II 5.3+0.4 5.1+0.1 Fraction III 6.0+0.6 5.6+0.3 1 The values (mean + SD, n = 3) represent enzyme activity, expressed as U/mg protein. Cells were exposed to 1.0 mg/ml green coffee bean extracts and fractions in culture medium for 24 h. Some cells were further treated with H 2O2 for 2 h. Cells incubated in culture medium were used as negative control. Cells treated with H 2O2 added culture medium were used as positive control.

186

Table 8 Description of genes that are differently expressed in Caco2 cells upon incubation with coffee and MRPs samples Gene Description Functional classification ALB Albumin Good binding capacity. Regulates the colloidal osmotic pressure of blood ALOX12 Arachidonate 12 Oxygenase and 14,15leukotriene A4 synthase lipoxygenase activity AOX1 Aldehyde oxidase 1 Produce H2O2 and under certain conditions, can catalyze the formation of superoxide. ATOX1 antioxidant protein 1 Could bind and deliver cytosolic copper to the homolog copper ATPase proteins. May be important in cellular antioxidant defense. CAT Catalase Protect cells from the toxic effects of H 2O2. CCS Copper chaperone for Delivers copper to copper zinc superoxide dismutase SOD (SOD) GPX2 Glutathione peroxidase Responsible for the majority of the glutathione 2 (gastrointestinal) dependent H 2O2reducing activity in the epithelium of the gastrointestinal tract. Could play a major role in protecting mammals from the toxicity of ingested organic hydroperoxides. iNOS inducible Nitric oxide Produces nitric oxide (NO) which is a messenger synthase molecule with diverse functions throughout the body. MBL2 Mannosebinding lectin Is capable of host defense against pathogens by (protein C) 2 activiating the classical complement pathway independently of the antibody. MPO Myeloperoxidase Part of the host defense system of polymorphonuclear leukocyte. MSRA Methionine sulfoxide Has an important function as a repair enzyme for reductase A proteins that have been inactivated by oxidation. MT3 Metallothionein 3 Binds heavy metals. Could play unique roles in homeostasis of the central nervous system and in the etiology of neuropathological disorders. Protective effect on DNA damage in response to reactive oxygen species (ROS). NCF2 Neutrophil cytosolic Required for activation of the latent NADPH factor 2 oxidase NME5 Nonmetastatic cells 5, Confers protection from cell death and alters the protein expressed in cellular levels of several antioxidant enzymes. May (nucleosidediphosphate play a role in kinase) spermiogenesis by increasing the ability of late stage spermatids to eliminate ROS. NOX5 NADPH oxidase, EF Generate superoxide and functions as a H+ channel hand calcium binding in a Ca 2+ –dependent manner. domain 5

187

Table 8 Description of genes that are differently expressed in Caco2 cells upon incubation with coffee and MRPs samples (continued) Gene Description Functional classification PRDX1 Peroxiredoxin 1 Reduce hydrogen peroxide and alkyl hydroperoxides PRDX4 Peroxiredoxin 4 Involved in redox regulation of the cell. Regulates the activation of NFкB in the cytosol by a modulation of IкBα phosphorylation. PRDX6 Peroxiredoxin 6 Involved in redox regulation of the cell. Reduce H2O2 and short chain organic, fatty acid, and phospholipids hydroperoxides. Play a role in the regulation of phospholipid turnover as well as in protection against oxidative injury. PREX1 Phosphatidylinositol Functions as a RAC guanine nucleotide exchange 3,4,5trisphosphate factor (GEF). dependent Rac exchange factor 1 PRG3 Proteoglycan 3 p53 responsive gene. Possesses cytotoxic and cytostimulatory activities. PRNP Prion protein Response to oxidative stress. Cellular copper ion homeostasis PTGS1 Prostaglandin May play an important role in regulating or endoperoxide synthase 1 promoting cell proliferation. (prostaglandin G/H synthase and cyclooxygenase) PXDN Peroxidasin homolog p53responsive gene SELS Selenoprotein S Involved in the degradation process of misfolded endoplasmic reticulum (ER) luminal proteins. SEPP1 Selenoprotein P, p53 responsive gene. Possesses cytotoxic and plasma,1 cytostimulatory activities. SRXN1 Sulfiredoxin 1 homolog Contributes to oxidative stress resistance by reducing cysteinesulfinic acid formed under exposure to oxidants in the peroxiredoxins PRDX1, PRDX2, PRDX3 and PRDX4. May catalyze the reduction in a multistep process by acting both as a specific phosphotransferase and a thioltransferase TXNRD1 Thioredoxin reductase 1 May possess glutaredoxin activity as well as thioredoxin reductase activity.

188

Table 9 Reducing power of AraSer MRPs fractions 1 Sample LR DR Fraction I 0.24+0.01 a 0.25+0.01 a Fraction II 0.22+0.02 a 0.24+0.03 a Fraction III 0.23+0.02 a 0.26+0.02 a Fraction IV 0.19+0.02 a 0.23+0.03 a 1 Reducing power (RP) values (mg CGA/g freeze dried samples) are the mean ± SD obtained for triplicate solutions measured in triplicate. LR = light roast; DR = dark roast.

Table 10 Effect of AraSer MRPs on the catalase activity in Caco2 cells 1 3 h 24 h 72 h Control 2.8+0.4 2.6+0.3 3.0+0.4 LR 2.2+0.3 2.0+0.6 2.8+0.4 DR 2.1+0.4 2.2+0.4 2.7+0.2 1 The values (mean + SD, n = 3) represent enzyme activity, expressed as U/mg protein. Cells were exposed to 1.0 mg/ml light roasted (LR) and dark roasted (DR) MRPs in culture medium for 3 h, 24 h and 72 h. Cells incubated in culture medium were used as control.

Table 11 Effect of AraSer MRPs on the glutathione reductase activity in Caco2 cells 1 3 h 24 h 72 h Control 17+1 18+2 19+4 LR 17+1 20+2 21+2 DR 17+1 18+2 19+2 1 The values (mean + SD, n = 3) represent enzyme activity, expressed as mU/mg protein. Cells were exposed to 1.0 mg/ml light roasted (LR) and dark roasted (DR) MRPs in culture medium for 3 h, 24 h and 72 h. Cells incubated in culture medium were used as control.

189 Table 12 Effect of AraSer MRPs extracts and fractions on Caco2 superoxide dismutase 1 activity with (+) or without () H 2O2 treatment

H 2O2 +H 2O2 Control 18+2 15+1 Defatted 20+2 17+3 Fraction II 17+3 15+1 Fraction III 20+1 15+1 1 The values (mean + SD, n = 3) represent enzyme activity, expressed as mU/mg protein. Cells were exposed to 1.0 mg/ml light roasted MRPs extracts and fractions in culture medium for 24 h. Some cells were further treated with H 2O2 for 2 h. Cells incubated in culture medium were used as negative control. Cells treated with H 2O2 added culture medium were used as positive control.

190